Interwar statistics, colonial demography, and the making of the twentieth-century refugee

Abstract

The historiography of the twentieth-century refugee typically unfolds as a tale of national displacement followed by international surrogate protection. This article challenges that narrative by reframing the modern refugee as an emerging category of statistics and demography. Focusing on the world’s first international refugee survey, which was led by former British colonial administrator John Hope Simpson in 1937–39, the article situates the attempt to count and classify refugees across borders within scientific debates on global population control and white resettlement. While refugees’ mobility initially eluded established parameters of national demographic measurement, Hope Simpson drew on precedents of census work and migration schemes within the British Empire to counter their unpredictability. Revealing how the tenet of colonial demography shaped mid-century views on the ‘refugee problem’, the article broadens the space of refugee history beyond nation states and international institutions and emphasizes the relevance of statistics in turning refugees into a global post-war category.

In 1938, former British colonial administrator Sir John Hope Simpson offered an unusual solution to the ‘refugee problem’ that had plagued international politics since the First World War. In a world of closing borders, Hope Simpson observed during a speech at the Royal Institute of International Affairs in London, the onward movement of refugees was increasingly stymied. Serving as the director of the world’s first international refugee survey, he criticized the League of Nations for not grasping the nature of the problem, despite the many decrees for refugee protection it had issued. ‘These legal definitions’, Hope Simpson admonished, ‘do not … give us really an idea of what a refugee is’. ¹ What could do so, however, were statistics that rendered refugees visible as a biopolitical aggregate. Relying on past census work, Hope Simpson predicted in his speech that Russian refugees would soon disappear, partially through old age and a high death rate and partially through successful integration in Europe and the Near East. ² Similarly, his proposed solution to the increasing number of refugees from Germany followed demographic supply and demand. ‘Now of all times when … we may look forward to a somewhat rapid decline in our population’, he announced to his British colleagues, ‘surely the time has come to welcome those who will counteract this fall’. ³

This vision of managing refugees neither accurately predicted nor prevented the mass displacement of the Second World War. Yet Hope Simpson’s survey, which was published in full a few months later, had one lasting legacy: in its attempt to clearly define the ‘refugee problem’ through a reliable count, the report offered a blueprint for our understanding of the modern refugee as a fixed category. Scholars of refugee history have largely traced the formalization of this category—from a vague nineteenth-century designator to an internationally recognized class—through the growing legal corpus that culminated in the landmark United Nations (UN) Convention relating to the Status of Refugees in 1951. ⁴ After decades of piecemeal legislation, the Convention comprehensively defined refugees by displacement resulting from political persecution, independent of nationality. ⁵ Although recent revisionist arguments have highlighted the politicized nature of how this legal protection is applied, they still foreground the assumed link between categorization and international protection. ⁶

This narrative, however, omits an alternative genealogy for the twentieth-century refugee that relies decidedly less on ideas of surrogate protection and human rights. Here, I highlight instead the parallel impulse for categorizing refugees emerging from interwar debates about global population. Drawing on archival sources from Britain, Switzerland, and the United States, I follow the international survey’s trajectory from its philanthropic inception across think tanks and international organizations, and I elucidate the scientific and political stakes inherent to refugee enumeration. As part of his survey, which was eagerly anticipated by national governments, Hope Simpson set out to define refugees as an abstract category coalescing around political persecution. Enumerating them across national borders, the survey marked a prime example of what Ian Hacking has termed ‘the creation of kinds among the masses’. ⁷

Following a process of collective category-making, statistics variously overlapped and competed with legal attempts at defining refugees as a new ‘kind of people’. Hope Simpson’s survey, for example, offered a draft wording that would feed into the UN Convention a decade later. ⁸ While historians today primarily cite his work as a descriptive data source, the broader intellectual link between Hope Simpson’s project and the UN legal framework was evident to contemporaries in the international civil service. John George Stoessinger, an employee of the short-lived International Refugee Organization (IRO)—the forerunner of the United Nations High Commissioner for Refugees (UNHCR)—and a former refugee himself, indicated as much in his reflection on post-war attempts to codify the refugee. Instead of referencing the milestone Convention of 1933 as a precursor, he suggested that the international community arrived at a definition that ‘would coincide with the classic one of Sir John Hope Simpson’ from the 1939 statistical survey. ⁹ What is now heralded as the ‘classic’ legal definition of the refugee was, in Stoessinger’s mind, itself drawn from another, quite differently framed ‘classic’.

Historically contextualizing the Hope Simpson survey elucidates a social-scientific legacy evident (and too often repressed) in the concept of the modern refugee. It also highlights the merits of employing global history as a framework for analysing refugee management: politically, interwar refugees may have appeared as an inter-national ‘problem’, but statistical calculations and the practical solutions based thereon emerged from the world of late imperial governance, transcended the borders of new nation states, and animated myriad stakeholders, including imperial officials and refugees themselves.

I first situate the survey within a broader discourse on population control that was riddled with interwar anxieties about the distribution of an ever-growing world population and intersected with economic debates. Refugees, alongside other ‘surplus populations’, inevitably became the object of calculation and classification on an international scale. However, as I show, refugees constituted a different kind of mobile population that resisted established parameters of demographic measurement. To solve this dilemma, Hope Simpson drew on his experience in colonial administration: on the one hand, refugees proved akin to colonial subjects that had long evaded an accurate count; but on the other, reimagined as colonists, they appeared integrable into existing, highly racialized migration schemes. Hope Simpson’s proposed solutions for the ‘refugee problem’, I thus argue, were both inspired by and fed into imperial schemes of population redistribution. Finally, I highlight the parallels between legal and scientific category-making by detailing how enumeration emerged in tandem with legal codification as a way to identify and solve the interwar ‘refugee problem’ and thus tangibly contributed to the making of the modern refugee as a category with global stakes and ambitions.

Shaping population(s): resources, territory, distribution

Scholars of refugee history typically frame efforts to ameliorate the interwar ‘refugee problem’ within a discourse on humanitarian assistance and the negotiation of human and minority rights in international law. The League’s attempts to find ‘durable solutions’, overseen by High Commissioner Fridtjof Nansen, included engineering population exchanges, advocating for mass naturalizations, and securing legal protection for stateless refugees. ¹⁰ These early relief activities have all received attention in scholarship, as they slowly established an international regime that guided national refugee policies but simultaneously reinforced national interests and often coincided with nation-building projects. ¹¹ As Claudena Skran has illustrated, the nascent refugee regime was widely accepted not only out of humanitarian concerns but also because it promised to decrease states’ individual responsibility—as well as the number of refugees in the main host countries—and proposed refugee resettlements based on national economic needs. ¹²

Although refugees gained international recognition and protection for the first time under the League in the 1920s, such relief efforts emerged partially because member states had a keen interest in a system of ‘burden-sharing’. With the repatriation of Russian refugees to Soviet-controlled territory constituting an impossible manoeuvre on both legal and humanitarian grounds, relief schemes quickly pivoted toward resettlement, which in turn necessitated reliable statistics. Nansen—who was not a career politician but had been trained in the natural sciences—proved a vocal proponent of gathering numeric knowledge. When assigned the post of High Commissioner for Russian Refugees in 1921, his first measure was to order a census to determine the exact size of the refugee population. ¹³ In subsequent discussions, legal questions figured as a secondary and largely pragmatic concern: they hindered the controlled cross-border movement of refugees to match them with vacant employment opportunities, and thus proved an impediment to ‘a more equitable distribution of Russian refugees’ across the continent. ¹⁴ While the League continued to debate refugees’ legal limbo in special committees and sought to redress it by issuing the so-called Nansen passports, the lack of rights was not the primary motivating factor for Nansen and his collaborators at the High Commission and the International Labour Office (ILO). Foregrounding operational concerns, international civil servants quite explicitly framed the rapidly multiplying national groups of refugees as an aggregate of economic valence.

This calculated approach to refugee management aligned with the League’s broader practice of collecting and disseminating information on a variety of international ‘problems’, including infectious diseases and crime. ¹⁵ Information gathering was guided by social-scientific principles, and statistics in particular appeared as a factual language that could circulate with ease across borders, even when it became increasingly harder for people and goods to do so. At the same time, the internationalist quest to document refugees in the 1920s and 1930s can be fruitfully read as embedded in broader scientific debates about global population control. Alison Bashford has demonstrated that these debates did not exclusively focus on reproductive technologies and ‘family planning’ to limit exploding birth rates: they also linked population to resources and territory in a neo-Malthusian manner. ¹⁶

While the League was interested in population questions, for example, it dealt with them primarily in its Economics Section. ¹⁷ Monitoring population developments across the globe, in 1926 the League’s Economic Intelligence Service began issuing the Statistical Year-Book of the League of Nations, which included estimates of every country’s population next to data on worldwide production and consumption. Although statistics were broken down by country and appeared as national figures, the League’s analysists subsequently correlated them with territory to make visible areas of high population density and urge international redistribution. From their point of view, southern and central Europe presented one of the worldwide hot spots that needed new territorial outlets to maintain economic stability.

Population redistribution was further debated at the World Population Conference in 1927, organized by birth-control advocate Margaret Sanger as the first formal get-together for the scientific study of global population. Held in Geneva, the conference featured a number of international civil servants and devoted one session to international migration. This session was headed by Albert Thomas, the director of the ILO, which had begun to statistically investigate the patterns of worldwide migration. ¹⁸ Following Thomas’ lead, the debate revolved around the imperative to manipulate migration patterns in the interests of population control. After all, as Thomas put it, ‘of all demographic phenomena, migration is the most susceptible to direct intervention and control’. ¹⁹ In discussion, a clear consensus emerged against legal immigration restrictions, such as those implemented by the USA, because they hindered the natural flow of ‘surplus population’ from Europe into less densely populated areas. By contrast, the assisted migration scheme advanced by the British government following the Empire Settlement Act of 1922 was referenced as an example of how to usefully manipulate migration patterns for political and economic gain. Similarly, the assistance of refugees, some of whom the ILO had attempted to resettle in South America through colonization schemes since the mid-1920s, was lauded as another ‘step of a practical kind’ in a policy of worldwide rebalancing. ²⁰ Thomas even went so far as to suggest that planned migration should form an essential part of any ‘rational’ population policy—so long as the movement was measured and controlled. ²¹

Quantifying the ‘refugee problem’: the conception of the Hope Simpson survey

Any attempt to manage and redistribute refugees as a ‘surplus population’, however, had to first establish how numerous it was. In 1936, the Rockefeller Foundation, an ardent supporter of research on population questions, began to circulate memos within its Social Science Section on the importance of a scientific study ‘relating to the present status of refugees’ that would provide an overview of the problem and ‘all possible solutions’. ²² The Foundation’s officers sought to gather a report they could deliver to the League’s projected Assembly in September 1938, presumably to lead the intergovernmental body to devote further resources. ²³

Their choice for executing such a study, which they deemed ‘of great practical as well as scientific interest’, fell upon the Royal Institute of International Affairs, also known as Chatham House, which had been founded in 1919 and quickly become one of the institutional centres of the nascent discipline of international relations. ²⁴ An Executive Committee grant from the Rockefeller Foundation over £6,000 in January 1937 (roughly £410,000 today) aimed to facilitate the completion of the survey by December 1938. ²⁵ Endowed with substantial financial investment—in stark contrast to Nansen’s perennially underfunded efforts to manage refugees at the League—Chatham House set out to assemble a small survey team to create an authoritative data set on European refugees. ²⁶

It was implicitly understood that the solution to the distinctly international ‘refugee problem’ would require equally international expertise. As recent scholarship has revealed, however, ‘international’ was often synonymous with ‘imperial’ at Chatham House, an institution steeped in the political mindset of British settler colonialism. ²⁷ No surprise, then, that for the survey’s directorship they courted Sir John Hope Simpson, a ‘man of distinction and experience’ when it came to population questions but also a loyal imperialist. ²⁸ Born in 1868, he trained for the Indian Civil Service and spent most of his professional life in India, first as a district magistrate and collector and then in higher administrative roles as Secretary and briefly Acting Chief Commissioner of the Andaman and Nicobar Islands. ²⁹ Having returned to the metropole during the First World War, he had a brief stint in Parliament, where he chaired the India Colonies Committee, but soon resumed his life abroad on a series of imperial and international assignments that involved the resettlement of displaced populations. ³⁰ He served as Vice-President of the Refugee Settlement Commission in Athens, which the League established after the 1920–1923 war between Greece and Turkey; was sent to Palestine by the British government for an investigation into immigration, land settlement, and development in 1930; was again recruited by the League to resettle refugees in China in the aftermath of the Yangtze River flood in 1931; and became Commissioner for Natural Resources in economically bankrupt Newfoundland until his formal retirement in 1936. ³¹ He thus came to the refugee survey with the long resume of a colonially trained technical expert—a clear antecedent, as Chris Courtney has suggested, to the post-war development officer’s profile. ³²

The scientific study Hope Simpson and a core team of sixteen investigators carried out between September 1937 and October 1938 produced a 640-page report. ³³ It brimmed with facts culled from published sources, reports from the League and various relief organizations, and ad-hoc visits investigators had made to individual countries. The report was roughly structured in three parts: first, an exhaustive summary of refugee movements in the early twentieth century, including their origins and causes; second, an overview of responses by governments as well as international and private organizations and a detailed description of the ‘machinery’ employed to bestow order on the chaos; and third, a comparatively shorter prognosis of future refugee movements with tentative policy recommendations.

Answering to a non-scientific advisory board, which included then-president of the League’s Nansen International Office for Refugees Michael Hansson, the survey was presented in a fluent, simple narrative. Each part featured plenty of prose and even basic legal analysis, drawn from special reports submitted by the investigators. Yet measures of progress and prediction were typically circumscribed through statistics: numbers of refugees registered, refugees assisted, refugees to be resettled. As the author of the final report, Hope Simpson also took great care to translate the political-economic concerns behind the ‘refugee problem’ into seemingly objective scientific terminology: in the laboratory of interwar Europe, the experimental challenge was to reach an optimal ‘dispersion’ of population through ‘infiltration’ and ‘absorption’ of refugees into respective host countries. To tackle said challenge, statistics seemed the logical tool: an unbiased determinant of the most desirable distribution pattern, it also served as a control mechanism to stabilize the ‘flow rate’ and establish ‘orderly emigration’ by recording, registering, and classifying refugees across borders.

Defying the laws of statistics: refugees as an object of quantitative analysis

Statistics as method merits close observation here because, in the framework of the refugee survey, its use was both traditional and experimental. Modern statistics had succeeded as an explanatory framework as part of a larger nineteenth-century intellectual turn away from a determinist, quasi-religious understanding of the universe towards the recognition of stochastic factors in the unfolding of social life. ³⁴ Borrowing from medicine, this new way of scientific reasoning set up normalcy and deviation as a central binary for understanding and controlling the social sphere, thereby making statistical laws not merely descriptive but prescriptive. ³⁵ Applied to demographic data, it could swiftly ‘diagnose’ phenomena of national under- and overpopulation. ³⁶ Further, as part of a wider state project of ‘legibility and simplification’, it offered a new way of knowing the body politic qualitatively. ³⁷ Autonomous statistical law often postulated a racialized normalcy and identified ‘problematic’ conduct by delineating particular subpopulations.

Twentieth-century international civil servants, too, threaded these normative dimensions throughout their practice of statistical inquiry, all the while presenting data collection and analysis as an objective, apolitical methodology. In the setting of the League, comparative statistics became crucial to generating shared international spaces, such as a stable world economy. ³⁸ Transnational numerical calculations also contributed to the idea of global population as an interconnected whole, in the vein of a living organism that could be scientifically ‘treated’ for under- or overgrowth.

Within this organism, however, refugees appeared as a foreign object, an odd pathology that seemed hardly curable through statistics. Designated as a special legal category by the League, they marked a novel variable in the established equation of population with yet unclear relations to other subunits. Not formally associated with any state—being either de jure stateless or de facto exiled from their home country—refugees could not feasibly be treated like any another ‘deviant’ subpopulation nor fully incorporated into national statistics. Still, they had to be measured in relation to nation-states in order to determine their present distribution and aggravating effects on population density hotspots.

For the 1939 survey, these conceptual problems reached deeper still because refugees were, quite literally, ‘moving targets’. ³⁹ For one, the statistical events traditionally used for measuring a given population—birth and death—proved ill-devised. Refugees’ numerical ‘death’ did not usually occur when they physically died, but rather when they categorically disappeared through either naturalization or exit of the national territory. In fact, their ‘death’ in one national census, when induced by leaving their host country, often entailed a ‘re-birth’ in another as they crossed national borders. In short, the vital markers of the refugee population were largely artificial because the statistical birth and death of a refugee rarely mapped onto the biological occurrences. This was, to an extent, by design: the refugee category was meant to be a temporary marker and not to last a lifetime, much less to be inherited across generations. For the purpose of tracking and projecting, however, it rendered vital statistics meaningless and made presence, not life, the prime factor to be counted.

Change over time within the refugee population was similarly registered through movement, not make-up, as expulsion rates replaced reproduction rates. Because their movement appeared unpredictable, driving refugees’ total towards zero emerged as the only way to safely return to a normal population count. After all, most states—as well as the League—sought to reduce the number of refugees, not to stabilize or increase it. Interwar refugees were thus constructed as a precarious category: their very right to statistical existence was constantly questioned as governments, international organizations, and humanitarians worked to effect a numerical ‘liquidation’. ⁴⁰ This in turn produced eerie discursive echoes of the persecution practices that brought them into existence as refugees in the first place. Yet, as Hope Simpson wrote in the preliminary version of the report, ‘absorption is a condition defying quantitative analysis’—i.e. it was hard to decide exactly when a refugee ceased to be one. ⁴¹

Recalcitrant masses: refugees, colonial subjects, and the limits of numerical surveillance

In addition to these conceptual conundrums, the survey’s team faced practical hurdles in assessing the quantitative dimension of the ‘refugee problem’. Throughout the final report, Hope Simpson acknowledged repeatedly that the numbers used in the survey were often mere estimates, considering that the usual surveillance tool of the identity card was largely nullified by refugees’ transitory legal status. The famed Nansen passport issued by the League had remedied this circumstance to a certain extent but was only reluctantly extended to groups other than Russian and Armenian refugees and thus remained unreliable for gauging the entire refugee population. ⁴² In attempting to count refugees, census takers instead depended on an aid network that encompassed both national governments and private relief organizations and was largely coordinated—but not streamlined—by the League.

Upon closer inspection, the refugee survey’s quantitative basis was derived from a variety of sources that did not employ the same units or standards. A 1921 estimate of Russian refugees in Finland was deemed to ‘probably exclude Jews and Ukrainians’. ⁴³ The figure for the foreign population in 1936 France was ‘thought to be less than the real total because of unwillingness to disclose foreign origin at a time of depression’. ⁴⁴ To calculate the potentiality of the Jewish population in Germany and Eastern Europe becoming refugees, the survey’s investigators added a seemingly arbitrary number, ‘probably inadequately’, to make up for hitherto unrecorded emigration. ⁴⁵ In the end, the report simply stated that ‘there can be no accurate statistics except of the assisted emigration’ from Europe. ⁴⁶

These shortcomings highlighted how international statistics, though ubiquitous by the mid twentieth-century, were not standardized methodologically across national borders. ⁴⁷ Strikingly, they also echoed perceived flaws in the adjacent realm of colonial statistics. Census-taking was an important tool of colonial control and social investigation widely applied across European empires. ⁴⁸ Judging by the writings of contemporaries, however, statistics often failed as an imperial science, and colonial populations remained ill-defined in the hands of demographers.

The same year that Hope Simpson and his team set out to quantify refugees, influential demographer Robert René Kuczynski, who had attended the 1927 Population Conference as a German representative but then emigrated to Britain, memorably complained that ‘the population statistics of most colonies are to-day in a condition similar … to that of the population statistics of most European countries 150 years ago’. ⁴⁹ According to Kuczynski, who worked as an advisor to the British Colonial Office, census-takers in the colonies were inconsistent in their calculation methods and their choice of categories, which made it impossible to track change from one census to the next, let alone make comparisons across different colonies. ⁵⁰ Further, Kuczynski reported that population estimates were broken down into flawed categories that did not correspond to the local social hierarchies colonial subjects used for self-identification. With birth and death registration often not compulsory or properly enforced, colonial censuses offered at best a ‘reasoned guess’ at the size and make-up of the population and produced knowledge that Kuczynski deemed ‘utterly inadequate’. ⁵¹

Given Kuczynski’s and Hope Simpson’s shared reference point, the British Empire, Hope Simpson was certainly aware of colonial demography’s problems. In fact, while he pursued the refugee survey, he maintained his involvement in a Chatham House working group that produced the lengthy 1937 report The Colonial Problem, which devoted an entire subsection to ‘The Population Problem’. ⁵² The parallels between the colonial and the refugee census were obvious. Like Kuczynski, Hope Simpson was attempting prognostics on top of statistical analysis and was engaging both cross-section and time-series data: comparing refugee groups in different countries at a certain moment while also predicting their change across time. ⁵³ And like colonial subjects, refugees often figured as non-national populations and—especially in the case of refugees arriving from ‘the East’—easily appeared as non-European to Hope Simpson’s contemporaries. ⁵⁴

Unsurprisingly, the recalcitrant subjects of Kuczynski’s and Hope Simpson’s respective inquiries also produced some of the same practical complications. As Kuczynski pointed out in his critique of colonial statistics, the quantitative assessment of the colonial population depended on the latter’s cooperation and therefore left an opening for resistance. Similarly, refugees’ agency posed a problem to their management, which Hope Simpson addressed in the final report with particular regard to Russian refugees. Expelled from their country of origin about a decade prior, they should have settled and ‘disappeared’ within statistics by the 1930s, he thought. But some resisted naturalization and ‘final absorption’ based on the politics of their national identification. Philosophizing about these often-stateless refugees from Russia, Hope Simpson wrote in the report: ‘on objective grounds their position was stabilized as repatriation became impossible; on subjective grounds their nostalgia stabilized them as refugees’. ⁵⁵ In France, Russian refugees retained their belief in an eventual return to their homeland even after twenty years of displacement. ‘To the outside observer this attitude cannot but seem somewhat pitiful’, Hope Simpson remarked. ⁵⁶ Persistent nostalgia and accordant political loyalties—a phenomenon extending beyond Russian refugees—was understandable but ultimately counterproductive in his eyes, as it meant that ‘in many ways the refugees themselves contribute to their insecurity by their political activity’. ⁵⁷ By hampering the natural flux inherent in the category, refugees actively turned a temporary into a permanent condition.

Yet the parallels between the two quantification projects ultimately reached their limits. Kuczynski successfully advocated for streamlining colonial census-taking and realized his aspiration to capture the entire British colonial population through one comprehensive census in the three-volume Demographic Survey of the British Empire in 1948. ⁵⁸ Hope Simpson, meanwhile, continued to struggle with his slippery object of study. Refugees’ simple presence appeared as a constant dilemma: unlike colonial subjects, they did not remain at a safe distance overseas and instead erased the imperial spatial separation by appearing in the midst of the ‘regular’ national population—indistinguishable unless further specified.

Making visible, letting disappear: imaginaries of refugee data

Following the example of colonial statistics, the first step in quantifying the interwar ‘refugee problem’ consisted of filtering out refugees from other migrants and the populace at large by instituting a centralized site for registration: the refugee camp. By default, the camp created a separate group that could be delineated and definitively counted. ⁵⁹ However, only those who landed inside its walls were recorded as legitimate refugees, while those who had sufficient means to settle outside of the camp were typically not. Nevertheless, Hope Simpson concluded in the survey that ‘temporary camps [are] a necessary stage in the process; there is no alternative’. ⁶⁰

Mass encampment of ‘suspect’ populations has a long prehistory within the British Empire and reveals the imperial origins of modern refugee management. ⁶¹ But interwar refugees’ temporary incarceration also replicated the power structures instantiated by the nationalist policies that caused their displacement, and the irony of rescuing would-be refugees from concentration camps (e.g. in Germany) only to install them in ‘humanitarian’ internment camps did not escape contemporaries. ⁶² In the refugee survey, Hope Simpson further had no qualms relying on the census data collected in concentration camps by Nazi Germany’s Racial Political Office to calculate the growing number of Jewish refugees. Despite his political disagreement with the German government, he posited in the report that if anything, their statistics were not detailed enough because, in their preoccupation with the camps, the German census-takers had not tracked emigration properly. ⁶³ Emphasizing the supposedly impartial nature of statistics, he suggested patiently waiting for Germany’s 1939-census to obtain more accurate, up-to-date numbers regarding Jewish emigration. ⁶⁴

At the same time, many among Hope Simpson’s own investigators had a personal interest in documenting and improving the fate of specific national groups. The investigators’ ‘special reports’ on individual countries constituted the study’s backbone and often linked back to local advocacy work. This was most obvious in the case of the lengthy reports on Russian refugees, which were drawn up by Russian exile jurists such as Arsène Stoupnitzky, Alexis Goldenweiser, and Jacques Rubinstein, who had emigrated westward after the revolution and were deeply invested in relief work. Similarly, in his report on refugees from Germany, social worker Salomon Adler-Rudel focused mainly on documentation by Jewish aid organizations—to the consternation of the survey’s main statistician Käthe Liepmann, who was tasked with synthesizing the data.

Having herself fled Berlin for Britain just years prior, Liepmann approached the ongoing expulsion of German refugees with particular scrutiny. The drawn-out nature of the persecution, which hinged on a gradual increase of political pressure rather than an immediate threat to life, made it hard to determine at which point a person officially became a refugee. ‘While we may think of the circumstances of the exodus of the Russian refugees as an abrupt decapitation,’ Liepmann drily remarked, ‘the persecution of non-political Jews in Germany resembles rather a gradual strangulation’. ⁶⁵ Further, since there was virtually no data on non-Jewish refugees fleeing from Nazi Germany, she worried, the international public would continue to perceive the persecution as a Jewish-only matter. ⁶⁶

Liepmann’s notes show that women, too, took part in the production of statistical knowledge about refugees, even if the published record often suggests otherwise. ⁶⁷ Rather than serving as a mere administrative assistant, Liepmann pursued highly skilled work and interrogated both origin and future use of the data that Hope Simpson presented as objective intelligence in the final report. This deep engagement with statistics led her to pursue a doctorate in sociology, which she received from the London School of Economics in 1942. For the refugee survey, however, Hope Simpson largely sidelined Liepmann’s critical commentary on who should be counted as a refugee, leaving her intellectual contributions—just as those of the ‘special investigators’—concealed by institutional politics.

Instead of cautioning readers against the unreliability of the existing data, the survey’s prognosis section focused on anticipating the exact increase in emigration as the Third Reich usurped neighbouring territories in the late 1930s. Camps served to make refugees visible as a population; yet the ultimate goal was to let them disappear statistically. If resettlement on new lands could not wholly eliminate the ‘refugee problem’, Hope Simpson figured, one could perhaps prevent the numerical growth of the category by integrating refugees into existing migrating populations and thereby convert their movement from an unknown threat into a calculated risk.

In legal terms, this could be as simple as switching out a label: ‘As soon as the immigrant is admitted’, Hope Simpson noted with a nod to the United States, where immigrants and refugees were not legally distinguished, ‘he ceases to be a refugee and therefore no specifically refugee problem exists’. ⁶⁸ Such a strategy for ‘absorption’ was a highly politicized matter, however, as he quickly learned. Upon attempting to obtain refugee data from across the Atlantic, he had a prolonged exchange with Joseph Chamberlain, a law professor at Columbia University and chairman of the National Coordinating Committee for Aid to Refugees and Emigrants Coming from Germany, which had been established in 1934. To Hope Simpson’s surprise, Chamberlain was reluctant to pass along data on refugees who had arrived in prior years. This hesitation, the correspondence suggests, was driven by concerns of data confidentiality, so as to not evoke any impression of ‘special treatment’ or legal trickery. ⁶⁹ Ultimately, Chamberlain not only rebuffed Hope Simpson’s requests and instructed the survey’s point person for the United States to do the same; he went so far as to contact the Rockefeller Foundation with concerns about Hope Simpson’s handling of refugee data. ⁷⁰ This manoeuvre worked: Hope Simpson promised to have the text cleared by both the Foundation and Chamberlain before publication, and the survey’s section on America shrank significantly from its early drafts. ⁷¹

Useful additions? Refugee resettlement between nation and empire

Still the fantasy of a categorical metamorphosis was maintained in the survey, for Hope Simpson devised an alternative scenario that would turn refugees into assets for the receiving countries. As the final report specified, they would have to be settled in a place with a favourable ratio of population to territory and suitable economic conditions. Here, Hope Simpson pursued an argumentation that harkened back to the political strategy of Albert Thomas. At the 1927 Population Conference, Thomas had praised the potential of controlled migration to stabilize markets by matching ‘excess labour’ to national economies in need. Notably, his approach also resembled the work of Nansen at the League, who had pushed for the orderly facilitation of refugee movements to reintegrate refugees into the nation-state framework.

To draw an explicit parallel between Nansen and Hope Simpson is tempting: after all, both were involved in the Greek-Turkish population exchange, the first major internationally organized refugee resettlement scheme. ⁷² As others have shown, Nansen played a decisive role in negotiating the 1923 agreement that infamously ‘unmix[ed] the populations of the Near East’ by swapping out Greek minorities in Turkey and Turkish minorities in Greece, which effectively led to ethnic cleansing across both countries. ⁷³ Hope Simpson formally became involved in overseeing the exchange as Vice-Commissioner of the Greek Refugee Settlement Commission in 1927. While he did not come up with the idea, he pursued his work enthusiastically and, in his reports, frequently justified the Commission’s work as a project of economic development and the creation of ‘a sturdy race of peasants resulting from a blend of all the elements of Hellenism’. ⁷⁴

Unlike Nansen’s primary mission of bolstering the new order of nation states, however, Hope Simpson’s work with refugees was decisively shaped by his experience as an imperial civil servant. This background neither escaped nor disturbed his employers at the League. As Charles Eddy, his superior on the Greek Refugee Settlement Commission, wrote to Director-General Eric Drummond, the service in India had provided Hope Simpson with ‘just the kind of experience which was required for grappling with the problems’ in Greece. ⁷⁵ But Hope Simpson’s investment in the colonial mission was not merely confined to the past: as confidential communication with the League shows, he almost resigned from his post on the Commission because he became a candidate for the governorship of Kenya in late 1929. ⁷⁶ Given this continued engagement with the British Colonial Office, Hope Simpson’s imperial assignments served as a relevant framework for the 1939 refugee survey in that they offered immediate insights into the resettlement and management of ethnically diverse groups.

This is true, above all, for the 1930 scientific report on Arab displacement in Palestine Hope Simpson devised on behalf of the British government. In 1929, just when the Greek Refugee Settlement Commission was wrapping up its work, Whitehall sent Hope Simpson to Palestine following the Arab population’s violent protests against Zionist territorial expansion. His expert assessment was meant to expand upon a report by the Shaw Commission, which intimated that given the limited land resources available in Palestine, the Arab population would not be properly protected as stipulated by the British Mandate if continuous Zionist territorial expansion was allowed or even encouraged. ⁷⁷ As someone familiar with resettlement and agricultural development, Hope Simpson was asked to weigh in and determine concrete steps to avoid further conflict. ⁷⁸

His findings, subsequently dubbed the Hope-Simpson Report, echoed the Shaw Commission in emphasizing the economically precarious state of the Arab population due to loss of land, although he blamed this fact on both Zionist expansion and Arab failure to modernize agriculture and other land-use techniques. As a solution, Hope Simpson suggested (a) restricting and increasingly policing Jewish immigration and (b) implementing a broad-scale development program, which would free up land by intensifying use and thus aid both parties. ⁷⁹ Although Whitehall deemed this plan too expensive, substantial parts of its general argument were picked up in the Passfield White Paper issued the same year. ⁸⁰

At the same time, Hope Simpson’s statistically driven plans faced resistance from parts of the affected populations. Given that his report called for definitive limits to the 1917 Balfour Declaration’s goal of establishing a Jewish national home, Zionist leaders were expectedly displeased. Their response deliberately did not focus on political arguments, however, but rather on the scientific validity of Hope Simpson’s study. In 1931, the Jewish Agency issued a scathing point-by-point rebuttal of the ‘hastily compiled’ and ‘very doubtful’ statistics employed in the report, effectively accusing its author of misreading the available data about Jewish settlers, using ill-defined categories, and performing ‘incomprehensible’ calculations. ⁸¹ In their reading, Hope Simpson had drawn large-scale conclusions about Arab displacement from unreliable data points and delivered ‘a verdict upon the work, the dreams, the last hope of a suffering people’. ⁸²

The Zionists’ blistering critique arose at least partially from the fact that the Hope-Simpson Report marked a direct prelude to Palestine’s much-anticipated 1931 census, which promised to translate body counts into political representation. ⁸³ But their take-down of Hope Simpson, though carefully formulated in the language of statistics, was a mismatch for the intellectual frame that the British administration employed when it came to Palestine. For the purpose of influencing the make-up of the upcoming census apparatus, Jewish statisticians rebutting Hope Simpson repeatedly invoked European conventions of national population statistics in order to pre-empt and invalidate the collection of data that would underpin arguments about immigration-induced Arab displacement. Yet British authorities had a different, decidedly non-national blueprint in mind: the Indian Census. ⁸⁴

The colonial frame of Hope Simpson’s Palestine report was also fuelled by his earlier work on the Indian Colonies Committee in the mid-1920s, where he had dealt with intra-imperial labour migration and questions about the extent to which Indian immigrants could drive—or hinder—economic progress among Africans in Kenya. After long deliberations over the available migration statistics, Hope Simpson’s committee had eventually recommended holding off on immigration restrictions for Indians, much to the dismay of the local white settler minority. ⁸⁵ The reasoning was that ‘the policy of the British Government was to do right and justice to white, brown and black alike’. ⁸⁶ In his Palestine report, despite recognizing that most Jewish newcomers were effectively refugees from Central and Eastern Europe, Hope Simpson similarly subsumed them under ‘imported labour’ in what he perceived as a colonial space. ⁸⁷ While affirming the perception of a qualitative difference between the Jewish and Arab populations, he maintained that the British authorities, under the Mandate, had to stand above both and ‘look upon the country as one unit’, ensuring that there was ‘no discrimination between the races which it contains’ in either politics or statistics. ⁸⁸ In both cases, Hope Simpson appealed to a seemingly horizontal diversity that, in practice, presumed a vertical stratification of colonial populations requiring constant and careful calculation to balance out. ⁸⁹

The main lesson Hope Simpson seems to have taken away from this legacy of managing heterogeneous populations was that resettlement, even when organized in consideration of land availability, labour needs, and racial fit, had absolute limits. A calculated policy of gradual ‘infiltration’, he reasoned, was thus necessary for European refugee resettlement in order not to disturb the fragile economic and racial equilibrium in the receiving countries. Utilizing this principle in the final report of the 1939 survey, he suggested first assessing the economic, political, and cultural ‘absorptive capacities’ of potential receiving countries, and then tailoring refugees towards these countries’ needs. Some refugees, Hope Simpson noted, ‘must be trained for other occupations before they can usefully be emigrated’. ⁹⁰ Engineering a worldwide resettlement scheme in which individual refugees were slowly added to countries that needed settlers for one reason or another would, in his eyes, have the most promising long-term results.

A ‘carefully regulated flow’: new settlers for the British Empire

Incidentally, one political space that was in need of a particular kind of settler in the 1930s was the British Empire. In its 1932 report, the Committee on Empire Migration re-emphasised ‘the desirability of a steady flow of migrants’ from the metropole to the overseas territories to evenly distribute the white population across the empire. Simultaneously worries abounded that there might not be enough white settlers to be redistributed due to Britain’s declining birth rate. ⁹¹ Predicated on this racialized discourse, imperial statistics became increasingly fixated on differential fertility between white and non-white populations, and interwar demographers compared fertility rates of populations in metropole and colonies in fear that white Britons would soon be outnumbered by non-European colonial subjects. None other than Kuczynski announced in 1937 that ‘fertility has decreased so much that the white population of the Empire is no longer reproducing itself’. ⁹² Referencing the historically favourable immigration rate of whites into the United States, he hinted that additional immigration into the empire could potentially solve the problem. This strategy was officially echoed by the British Oversea Settlement Board, which in a 1938 report advocated for ‘the admission of a carefully regulated flow of foreign immigrants of assimilable type … whether of Northern or other European extraction’ in order to bolster and redistribute the white population across the empire. ⁹³

Might European refugees, whose ethnicity was never openly discussed in the purportedly neutral language of social-scientific inquiry, help remedy this precarious demographic state? Neither Kuczynski nor the Oversea Settlement Board went as far as to state this outright, and Hope Simpson did not explicitly acknowledge such plans in his refugee survey either. He did, however, allude to that possibility elsewhere. In 1939, the year the survey was published, Hope Simpson reviewed White Settlers in the Tropics, a study by Australian geographer Grenfell Price that analysed the challenges faced by newly-arriving white settlers and ‘near-whites of south European descent’: physical maladaptation to the tropical environment, and competition with ‘inferior populations’ of lower racial and economic status. ⁹⁴ In his assessment, Hope Simpson announced that the positive experience of Italian immigrants in Northern Australia ‘should be of great value in the investigation of areas of possible settlement for refugees’. ⁹⁵ Not only were interwar refugees assumed to be at least ‘near-white’, then, they also appeared to rank as more ‘civilized’ than colonial subjects. They could act as a buffer between non-white natives and forced labourers on the one hand and ‘truly’ white settlers on the other. Indeed, as Jochen Lingelbach has shown, when British colonies in Africa became an unexpected refuge for some of the displaced in the mid-1940s, the local colonial societies tended to treat them as ‘subaltern whites’. ⁹⁶

The solution Hope Simpson ultimately favoured for the ‘refugee problem’ relied upon exactly such an imperial scheme of turning refugees into settlers. Conspicuously, it drew on the nineteenth-century practice of rerouting incoming foreigners from the metropole to the colonies in order to provide long-term refuge without straining British social relations at home. ⁹⁷ Yet if those earlier transmigrations aimed, as Caroline Shaw has argued, at providing a real alternative to persecution and ‘overseas outlets’ had proved instrumental in garnering public support for large-scale humanitarian assistance, by the mid-twentieth century the motivation had shifted. Refuge and asylum no longer figured as moral imperatives in the technocratic debates about refugees. Instead, resettlement plans followed the ostensibly dispassionate algebra of population distribution.

In Hope Simpson’s mind, the advantages of overseas resettlement by ‘infiltration’ were clear and echoed official government policy for intra-imperial migration. ⁹⁸ For one, it would allow the use of existing infrastructure and not necessitate new ‘empty land’, which was costly and increasingly hard to find. ⁹⁹ ‘Colonizing settlement in an under-developed territory’, he noted, while not to be ruled out entirely, ‘is necessarily a gradual and slow process, which must be thought of in generations’. ¹⁰⁰ The largely unsuccessful ILO schemes to resettle Russians in South America in the 1920s had shown as much. ¹⁰¹ Second, an imperial scheme could proverbially kill two birds with one stone by avoiding American financial and institutional support and solving the British Empire’s demographic dilemma. Transforming refugees into settlers promised to evenly distribute and boost white—or ‘near-white’—populations across Britain’s colonies and dominions. As Hope Simpson asserted at a meeting of the Overseas League in 1939, there were ‘certain parts of the Empire which unless they consented to such immigration might have to face the danger of immigration from Oriental countries’. ¹⁰² When it came to strategies for solving the ‘refugee problem’, it clearly appeared useful to think of refugees as a category more akin to colonizers than colonized. ¹⁰³ Plus, privileging gradual ‘infiltration’ over ad hoc schemes and integration into existing colonies over new land settlements reinforced existing power relations and guaranteed Europe’s status quo.

Later that same year, British journalist Norman Angell and activist Dorothy Buxton drew on Hope Simpson’s report to make a more impassioned appeal to integrate refugees into the British Empire. In their manifesto, they argued that continued British isolationism would only prolong economic stagnation and hasten the demise of the white empire. Whether one pitied them or not, European refugees were the imperial population’s saving grace. In contrast to Hope Simpson, Angell and Buxton also pushed beyond measured infiltration and considered mass resettlement in a ‘vast undeveloped country’ like Australia a viable option—especially since refugees, given their lack of options, would likely have a higher ‘readiness to face hardship’ than ordinary settlers. ¹⁰⁴ ‘Can a general self-interest move us where humanity fails?’, the authors asked on the eve of the Second World War, urging for a reversal of immigration policies across the Commonwealth. ‘A few years may entirely change the scene, and an eager competition may develop for such refugees as survive. By that time the best will have perished.’ ¹⁰⁵ To derive a tangible benefit from the mass displacement, they reasoned, assistance should be offered before wartime effects had permanently altered the utility of the European refugee population.

Law and numerical order: resettling and redefining the post-war refugee

Although never enacted on the exact terms and scale either Hope Simpson or Angell and Buxton had wished for, ideas about imperial population redistribution echoed throughout resettlement policies for the millions of European refugees and Displaced Persons (DPs) in the aftermath of the Second World War. ¹⁰⁶ Next to physical fitness, perceived racial characteristics among post-war refugees—many of whom were from Eastern Europe and did not want to be repatriated to countries now under Soviet rule—determined their migration routes into and within the Commonwealth. Between 1947 and 1952, the Australian government took in around 170,000 DPs, who made their way across the ocean with the help of the IRO and were ‘preferably fair-skinned’ and easily transformed into ‘assimilable “white” migrants to make up for a disappointing absence of postwar British settlers’. ¹⁰⁷ British foreign labour recruitment in the late 1940s under the European Volunteer Workers (EVW) scheme vied for those same refugees to cross the Channel and replace the ‘British stock’ that the Home Office intended to send out to the dominions in order to rebolster imperial ties. ¹⁰⁸ The Royal Commission on Population, tasked with conducting a grand study of British population trends in 1944, concluded in its final report in 1949 that ‘a net inward balance of migration of 140,000 young adults’ would be beneficial for counteracting both the general decline in birth rate and the simultaneous emigration schemes required for ‘maintaining and strengthening the British element in the Commonwealth’. ¹⁰⁹ Though citing the EVWs favourably, the report also alerted readers of the dwindling supply of such easily assimilable ‘good human stock’. ¹¹⁰ In no uncertain terms, the calculated intake of ‘near-white’ continental refugees was deemed preferable to the immigration of non-white British subjects from Asia or Africa. ¹¹¹

But resettlement schemes were not the only post-war feature nascent in Hope Simpson’s report. As mentioned in the beginning of this article, his definition of the refugee seamlessly blended into international law. Yet this definition is best understood as a product of, and not the basis for, the statistical survey. For one, Hope Simpson evidently sought to go beyond the League’s efforts to protect those who could not return to their country of origin because they had lost their citizenship. By the mid-1930s, the League officially recognized only Russians, Armenians, Greeks, Bulgarians, Turks, Assyrians, Germans, and Saar expellees as refugees, and these were considered strictly by nationality—a fact that other contemporaries, like American journalist Dorothy Thompson, critiqued as well. ¹¹² While Thompson saw the League’s narrow legal definition as constraining refugee statistics, however, Hope Simpson treated numbers as potentially untethered from the ‘arbitrary limits’ set by political expediency. ¹¹³ Although yielding to the practicality of focusing on those refugees ‘in whom the League [was] already interested’—namely European refugees scattered across Europe, the United States, and the Far East—his survey posited a more universal rationale: ¹¹⁴

The essential quality of a refugee … may be said to be that he has left his country of regular residence, of which he may or may not be a national, as a result of political events in that country which render his continued residence impossible or intolerable, and has taken refuge in another country, or, if already absent from his home, is unwilling or unable to return, without danger to life or liberty, as a direct consequence of the political conditions existing there. ¹¹⁵

Because he approached refugees as a statistical entity, it was clear to Hope Simpson that no matter their legal status, all refugees were to be considered part of the same unstable subpopulation that, by its very existence, challenged the assessment and distribution of global population.

The resulting category, international as it was, still largely functioned via negative definition. Hope Simpson distinguished refugees from minorities, which were covered by the League’s decrees and merely potential refugees, and from stateless persons, which could but did not have to be synonymous with refugees and called for ‘separate investigation’. ¹¹⁶ Interestingly, the survey also excluded groups forced into flight by natural disasters—such as those displaced by the 1931 Yangtze River flood. Hope Simpson acknowledged them as refugees but did not include them in the enumeration of the final report, leaving it open to interpretation whether environmental causes rendered refugeedom more ‘natural’, or whether he thought that these refugees, neither European nor ‘near-white’, were useless for imperial plans.

Finally, the report specified that the modern refugee ‘is distinguished from the ordinary alien or migrant in that he has left his former territory because of political events there, not because of economic conditions or because of the economic attractions of another territory’, thus laying the groundwork for our contemporary categorical divide between ‘political refugee’ and ‘economic migrant’. ¹¹⁷ But in contrast to today’s debates focused on the supposed threat of economic competition, the problem for Hope Simpson was not that individuals migrated according to market principles of supply and demand; it was precisely that refugees were on the move without any consideration for them. As the later sections of the survey show, refugees according to his definition may not have been generated by economic difficulties, but they could certainly resolve them if transformed into calculable economic actors. This attitude, too, resurfaced in post-war policy: under the helm of the newly formed Council of Europe, a nascent European administration emphasized the need for a more even distribution of manpower as well as for overseas outlets to resettle its ‘surplus elements’, of which they considered refugees to be a key part. ¹¹⁸ Indeed, there is evidence that the framing of refugees as a statistically measurable labour force continues into present-day governance. ¹¹⁹

By eliminating what refugees were not, Hope Simpson derived a definition that was both more specific and more inclusive than the League’s legal agreements. Despite its pretence to universality, however, this new refugee category was neither universal nor neutral. Its exclusive focus on political persecution echoed nineteenth-century British discourses on foreign refugees but disposed of the impression of those seeking refuge as ‘paragons of liberal virtue’. ¹²⁰ Twentieth-century debates were led in an entirely different register: what had once been honourable revolutionaries and hard workers were now idle ‘elements’ that needed to be actively managed. Nevertheless, the definition demarcated a coherent, sufficiently abstract category that statisticians could use for calculation and prediction.

This was necessary, Hope Simpson thought, in order to understand and solve the actual problem. ‘The invaluable services rendered by the League’s protection in the past and its indispensability for the present and immediate future’, he noted, ‘should not obscure the fact that it is concerned only with solution of the interim problems of the refugees, not with a final solution’. ¹²¹ The 1939 survey, by contrast, was a project of abstraction, using numbers as a means to visualize the collective that made up the refugee as a new ‘kind’. In the mind of Hope Simpson and his expert contemporaries, statistics took over precisely where international law could not reach. It was a way of truly ‘knowing’ and efficiently managing refugees.

Notwithstanding this competitive framing, statistical and legal efforts often worked in tandem rather than against one another to address the interwar ‘refugee problem’. ¹²² The survey’s final report proposed an ostensibly universal language that proved useful to subsequent legal efforts, including the 1951 Refugee Convention. In effect, taking statistics as the starting point helps explain some of the latter’s limitations in scope. This becomes clear, for example, from debates led at the UN Conference of Plenipotentiaries on the Status of Refugees and Stateless Persons that took place in 1951 to finalize the exact wording of the Refugee Convention. On November 26, 1951, national representatives discussed the definition of the term ‘refugee’ in Article 1 of the Convention and debated whether the qualifier ‘in Europe’ should be included to limit its scope geographically. ¹²³ Proponents of this qualifier repeatedly brought up the fact that the total number of refugees from beyond Europe—which had been outside of the scope of Hope Simpson’s survey—was simply not known, and that using a mere guess as the basis for legislation would be, as French representative Robert Rochefort put it, comparable to issuing a ‘blank cheque’. ¹²⁴ Only the numbered refugee was a known quantity and could therefore be guaranteed rights and assistance. In the 1951 Refugee Convention, the law followed statistics in rendering visible the twentieth-century refugee as a clearly delineated population—one that was, in the end, deemed distinct from colonial subjects, who were explicitly excluded from protection. ¹²⁵

Conclusion

Rereading twentieth-century refugee history from the vantage point of statistics and demography highlights historical conjunctures that legalistic terms gloss over more easily. It reveals that the task of quantifying refugees was integral to resettlement schemes—and, in the frantic search for new destinations, often linked back to colonial governance. To a former colonial administrator like Hope Simpson, the imperial model offered an alternative blueprint to the often inexperienced efforts of the League for governing disparate populations. Recent scholarly efforts illustrate the intertwining of post-war refugee politics and decolonization; yet the Hope Simpson survey shows that this link between the imperial and the international was forged even earlier. ¹²⁶ Schemes of resettlement and relief did not only advance or conflict with colonial interests in a post-colonial world. The very techniques of refugee management emerged from the afterlife of colonial expert knowledge. ¹²⁷

At the same time, the intertwined histories of studying and managing refugees betray broader ideas about international order and the world’s increasing interconnectedness in the mid-twentieth century. Taking a global historical perspective, as this article demonstrates, brings new historical actors into view, recontextualizes known ones, and shows their mutual embeddedness in transnational knowledge networks. Scaling up also reveals the global vision many of these actors pursued, as far-flung resettlement schemes rested on a conception of the worldwide population as an integrated whole that could be brought into equilibrium by means of controlled human mobility. More so than interwar law, which was generally articulated in deference to the nation state, international civil servants and experts actively deployed statistics as a means to universalise the notion of the refugee.

The particular case of the Hope Simpson survey also draws attention to how the mechanisms of counting and categorizing refugees were deeply intertwined. Following the long history of statistics as a social-scientific tool to identify and track deviant subpopulations, the choice of this method for refugee management constituted an attempt at managing a collective deemed ‘out of place’ or, in the case of stateless individuals, ‘out of order’. Understanding the Hope Simpson survey as an authoritative data source and lauding it as the ‘earliest modern study on refugees’—as has been the tendency in contemporary scholarship—is thus misguided in two ways. ¹²⁸ The analysis above has shown that the data it compiled was often confusing and unreliable. But more importantly, the survey does not mark a study on refugees, since this would presuppose that the category already existed when Hope Simpson and his colleagues were pursuing their expertise. Rather, in conjunction with practices of confinement and compulsory resettlement, their statistics contributed to generating refugeedom as a social reality. ¹²⁹

‘If new modes of description come into being’, Ian Hacking has posited, ‘new possibilities for action come into being in consequence’, which proves as true for scientific labelling as it does for legal categories. ¹³⁰ By the 1940s, the quest for refugee tallies had become a significant concern among international civil servants, relief workers, and social scientists, as evidenced by the growing corpus of refugee censuses. ¹³¹ Importantly, some of the most widely cited surveys were pursued by (former) refugees themselves. ¹³² Together with the ‘special reports’ that fed into Hope Simpson’s survey, such publications raise questions about representation and empowerment in refugee quantification, and about its intertwining with other political projects, including Zionism and anti-colonial activism. Exploring these dimensions, I suggest, can contribute to answering the recent call for ‘writing refugeedom into refugee history’. ¹³³ Instead of using surveys as factual sources to depict the scope of twentieth-century displacement, it is time to unravel their hidden calculations and ask what statistical practices actually promised, and sometimes delivered, to those in the field.