From Origins to Frontier: A History of Demography, Migration, and Urban Futures

How industrializing Europe's fertility and mortality decline, the great urban migrations of the twentieth century, and the UN's 1946 counting apparatus produced the population science that now tries to forecast the century ahead.

Counting a nation: a mid-century population tabulator processes census cards, one card frozen mid-feed. — Image prompt and art direction by Brecht Corbeel; image generated to that direction.

Abstract

Modern population science did not begin with a model; it began with parish registers, census counts, and the puzzle of why death rates fell decades before birth rates did. This article traces that puzzle through Sweden's and England's nineteenth-century vital records, follows the mass movements that emptied countrysides and filled cities on three continents, and ends at the founding of the UN Population Division in 1946, whose cohort-component projection method still structures how the world argues about aging, migration, and urban futures. Fact, vendor claim, analysis, scenario, and prediction are kept explicitly separate throughout.

Modern population science did not begin with a model. It began with a register: a parish clerk recording births and deaths in a bound ledger, one entry at a time, for reasons that had nothing to do with theory. The theory came later, forced into existence by a pattern the registers themselves made undeniable — that death rates in industrializing Europe fell for decades before birth rates followed, opening a gap in which populations grew faster than anyone alive had experience of. That gap is the demographic transition. Tracing how it was first observed, how mass migration reorganized the resulting populations across cities and continents, and how a UN agency founded in 1946 turned scattered national counts into a shared global forecasting apparatus, is the subject of this history.

This history matters for a reason beyond antiquarian interest. Every contemporary argument about population aging, migration pressure, or urban housing shortfalls implicitly borrows a vocabulary and a set of measurement conventions built between roughly 1880 and 1950 — parish registers giving way to national civil registration, national civil registration giving way to systematic census-taking, and census-taking giving way to an internationally coordinated statistical apparatus housed at the United Nations. Understanding which parts of that inheritance are hard-won empirical fact, which parts are theoretical generalization from a limited historical base, and which parts are forward-looking scenario work is not a philosophical nicety. It determines whether a headline claim about “the world’s demographic future” should be read as a report of something that has already happened, a considered extrapolation with named assumptions, or a bet.

The registers before the theory

Vital registration in parts of Europe predates the nineteenth century, but it is the systematic, near-continuous parish and civil registers of Sweden that give demographers their longest clean run of data on the transition as it happened. Sweden’s life expectancy at birth rose from about 32.9 years in the mid-eighteenth century to 43.9 years a century later, and to 58.8 years by the start of the twentieth century — a gain driven overwhelmingly by falling infant and child mortality, itself tied to improving sanitation and nutrition [4]. This is a fact recorded directly in the registers, not an inference from a model.

What makes Sweden’s case pedagogically central is the lag it exposes. Fertility did not move with mortality. Total fertility held in a stable band of roughly four to five births per woman — around eight per married woman — through most of the nineteenth century, and only began a sustained decline around 1880, gathering pace after 1900 [4]. For a full century, Sweden’s population grew rapidly not because families were choosing to have more children but because fewer of the children already being born were dying. That is the demographic transition in its rawest empirical form, decades before anyone named it: mortality decline, then a fertility decline lagged by a generation or more, with population growth concentrated in the gap between the two curves.

England’s version of the same story shows up not in vital registers but in the national census, begun in 1801. At that first count, roughly 30 percent of the English population already lived in urban centers of 2,500 or more; by the 1851 census, over half did — the first time in English history that town-dwellers outnumbered the rural population [8]. Using the stricter threshold of towns above 10,000 inhabitants, the share rises from about one-fifth in 1801 to two-fifths by 1851 [8]. Manchester’s population grew from 75,000 in 1801 to 351,000 by 1871; Liverpool grew from roughly 5,000 in 1700 to 83,000 in 1801 and 376,000 by 1851 [8]. These are census facts, not projections, and they describe a rural-to-urban reallocation of population that ran on a similar timetable to the mortality and fertility shifts documented in Sweden, though England’s industrial labor demand pulled people into cities faster and earlier than the transition alone would have.

It is worth being precise about what these two national cases can and cannot support. They establish, as documented fact, that mortality decline preceded fertility decline in at least two industrializing European populations, and that urbanization in England ran on a broadly parallel clock. They do not by themselves establish that this sequence is a universal law of development — a claim later theorists would make, and a claim this article treats separately, below, as analysis rather than fact.

Naming the pattern: demographic transition theory

The term “demographic transition” belongs to Frank Notestein, writing in 1945. Notestein, who a year later would become the first director of the newly created UN Population Division, published “Population: The Long View” as a chapter in Food for the World, refashioning an earlier model by Warren Thompson (1929) into a staged account of how populations move from high mortality and high fertility, through a transitional period of rapid growth, to low mortality and low fertility [2]. Notestein’s own three-part typology — “incipient decline,” “transitional growth,” and “high growth potential” — was explicitly built to explain the century of data just described: mortality falls first, driven by “technological, social, economic, and political developments”; fertility follows only after cultural and institutional norms around family size adjust to the new mortality regime [2, 7].

This is where fact ends and theory begins, and the distinction matters for how the framework should be used today. That mortality preceded fertility decline in nineteenth-century Sweden and comparable European populations is a documented empirical regularity. That this sequence constitutes a general law every population must pass through, in the same order, for the same reasons, is Notestein’s analytical generalization from that regularity — a generalization later demographers have qualified heavily, since twentieth-century transitions in Asia and Latin America ran on compressed timelines with different institutional drivers, and some populations show fertility and mortality declining together rather than in strict sequence. The theory is real, influential, and grounded in real data; it is not a physical law, and this article treats it as the analytical framework it is, not as an unconditional fact about all human populations.

The great movements: migration as a documented, not modeled, event

Alongside the transition, the same industrializing decades produced the first systematic attempt to find regularities in migration itself. E. G. Ravenstein, working from British census returns for 1871 and 1881 and later extending his data to continental Europe, the United States, and Canada, published “The Laws of Migration” in two parts, in 1885 and a second paper in 1889 [5]. Ravenstein’s proposed regularities — that most migrants move short distances, that migration proceeds in absorption waves as people fill vacancies left by others moving further, that migration flows generate compensating counter-flows, that natives of towns are less migratory than the rural-born, and several more — were among the first attempts to treat migration as a population process with statistical structure, rather than as a set of individual biographies. Several of Ravenstein’s regularities have held up reasonably well under later scrutiny; others have been revised or restricted to certain contexts. Either way, the 1889 paper is a primary source: it is Ravenstein’s own tabulation of census-derived flows, not a later paraphrase of it.

The migration event with the deepest documentary record in American history is the Great Migration, which the National Archives dates from 1910 to 1970: roughly six million Black Americans left the rural South for northern, midwestern, and western cities, one of the largest internal movements of people in U.S. history [3]. The Archives’ own periodization identifies two distinct waves — a first, 1910 to 1940, moving about two million people into northern industrial cities, and a second, 1940 to 1970, moving at least five million more, this time including many migrants who already had urban skills from Southern cities [3]. By 1970, nearly half of all Black Americans lived outside the South, concentrated in northern cities [3]. The documented drivers were escape from racial violence and Jim Crow legal subordination, and pursuit of industrial-era economic and educational opportunity [3] — a vendor-neutral characterization drawn directly from the National Archives’ own account, not an inference this article is making independently.

It is important to be exact about what the Great Migration and Ravenstein’s laws jointly demonstrate and what they do not. They demonstrate that mass internal migration, driven by a specific, documentable mix of push (violence, legal subordination, agricultural mechanization displacing sharecroppers) and pull (industrial wage labor, urban institutions) factors, reorganized national urban populations at a scale visible in a single national census-to-census interval. They do not demonstrate that all subsequent migrations follow the same causal mix — every migration wave since has required its own documentary reconstruction, and this article does not extend the Great Migration’s specific causal account to other movements not examined here.

Counting the world: the UN Population Division, 1946 onward

The institutional machinery that eventually tied national demographic and migration records into a single global forecasting system began in 1946. The Population Commission was established by the UN Economic and Social Council in resolution 3(III) of 3 October 1946, and the Population Division was created as its secretariat, with Frank Notestein — the same demographer who had named the transition a year earlier — as its first director [1, 9]. This is a documented institutional fact, not an inference: the person who supplied the theoretical vocabulary for describing population change became, within a year, the head of the body responsible for measuring and projecting it for every UN member state.

The Division’s founding mandate combined two functions that are easy to conflate but analytically distinct: producing internationally comparable demographic estimates — counts of what has already happened, built from national censuses, vital registration, and surveys — and producing demographic projections, forward-looking scenarios built on assumptions about future fertility, mortality, and migration [1]. The Commission’s second session identified the need for population estimates and forecasts as a priority almost immediately after the Division’s creation [1]. The methodology that resulted, and that still underlies the World Population Prospects series today, is the cohort-component method: a population is broken into age and sex cohorts, each cohort is aged forward one year at a time, and cohort-specific assumptions about fertility, mortality, and net migration are applied to produce the next year’s population, cohort by cohort. It is, mechanically, exactly what a hand-cranked adding machine and a shelf of bound life tables were built to do before electronic computing existed to do it faster — the historical equipment differs, the arithmetic structure does not.

This is a genuine methodological throughline worth stating plainly as analysis rather than overclaiming as unchanged fact: the cohort-component approach the UN Population Division formalized in the late 1940s is the same structural method — age cohorts aged forward under explicit fertility, mortality, and migration assumptions — used in the UN’s current World Population Prospects releases. What has changed is the quality and coverage of the underlying vital registration and census data feeding it, the computational speed at which cohorts can be aged forward, and the sophistication of the uncertainty ranges attached to the assumptions, not the basic cohort-aging logic itself.

From counted transition to counted urban future

The transition and the migrations it helped drive converged, over the twentieth century, on a single observable outcome: the world crossed from a predominantly rural to a predominantly urban species. The UN’s 2018 World Urbanization Prospects revision put the 2018 urban share of world population at 55 percent, up from a much smaller share in 1950, and projected a rise to 68 percent — roughly two-thirds of humanity — by 2050 [6]. That 2050 figure is explicitly a UN projection, not a fact yet observed, and it carries the same cohort-component logic traced above: an assumption about future fertility, mortality, and rural-to-urban migration rates, aged forward from the most recent verified census data. The 55 percent 2018 figure, by contrast, is an estimate drawn from actually reported national data and stands on firmer ground as a description of an already- observed state.

Reading these two figures side by side illustrates the fact/projection distinction this history has tried to maintain throughout. England’s 1851 urban majority is a census fact, verified after the event. The UN’s 55-percent 2018 global urban share is an estimate, reconciled across many national censuses of varying vintage and quality, carrying real but bounded uncertainty. The 68-percent 2050 figure is a projection, conditional on assumptions about fertility and migration trends holding roughly as the Division’s demographers currently expect — an explicit scenario, with a stated horizon (2050), stated assumptions (continued urban-rural fertility differentials, continued net rural-to- urban migration in currently low-urbanization regions), and an implicit disconfirmation condition: if fertility in currently high-fertility, low-urbanization regions falls faster than assumed, or if rural-to-urban migration slows for reasons not yet visible in the data (climate-driven rural livelihoods loss operating in the opposite direction, for instance, by displacing rural populations toward secondary cities faster than the model assumes, or urban housing and labor-market saturation slowing the pull), the 68-percent figure would need revision. The UN’s own projection rounds have revised prior rounds’ urban-share numbers before, which is itself evidence that the figure is appropriately treated as a scenario rather than a settled fact.

What this history does and does not license

Three genuinely distinct kinds of claim have been made in this account and it is worth separating them one final time. First, fact: Swedish mortality and fertility data, English census urbanization figures, the Great Migration’s documented volumes and dates, and the UN Population Division’s 1946 founding are all matters of historical record, drawn from primary registers, census counts, and the Division’s own institutional history. Second, analysis: Notestein’s demographic transition theory is a documented, historically important generalization from that record, not itself a physical law, and its applicability to twentieth-century transitions outside industrializing Europe requires separate justification this article has not attempted to supply in full. Third, scenario: the UN’s 2050 urban- share figure, and by extension any forward-looking claim in contemporary demography about aging populations, climate-driven mobility, or future urban housing demand, inherits the same cohort- component structure Notestein’s successors built in the late 1940s — a structure that produces rigorous, auditable, revisable conditional forecasts, not prophecy. The history of demography is, in this sense, also a history of learning to say clearly which of the three one is doing at any given moment — a discipline the parish clerks who kept the first registers could not have anticipated needing, and one the field still practices imperfectly today.

The equipment changed at every stage of this history, and the succession of instruments is itself a useful timeline. Iron-gall ink and ruled ledger columns were the entire technology of demography for most of the nineteenth century: a clerk, a quill, and a register bound to survive a century of handling. Electromechanical punch-card tabulation, developed for the 1890 U.S. census and matured through the 1940 count, let a national statistical office cross-tabulate age, sex, and location at a speed no clerk could match by hand, which is precisely what made cohort-level analysis of a whole national population practical rather than merely theoretical. Hand-cranked adding machines and bound cohort life tables were the working desk tools of the UN Population Division’s earliest projection rounds, before electronic computing took over the cohort-aging arithmetic in later decades. None of these instruments changed what demography was trying to measure — births, deaths, and moves, sorted by age and place — only how much of it could be measured, how quickly, and at what geographic scale. That continuity of purpose across a discontinuity of instrument is worth remembering whenever a new data source (satellite-derived settlement maps, mobile-phone mobility traces, administrative migration records) is proposed as a successor to the census: each past instrument earned its place by answering the same handful of demographic questions faster or at wider coverage, not by asking new ones, and each new source should be judged by the same standard rather than assumed superior on novelty alone.

Figure 1. Before the transition had a name, it had a register: birth and death entries recorded by hand, one parish at a time. — Image prompt and art direction by Brecht Corbeel; image generated to that direction.

Figure 2. Migration left a paper trail: manifests, route pins, and headcounts that later became Ravenstein's laws. — Image prompt and art direction by Brecht Corbeel; image generated to that direction.

Figure 3. The 1940 U.S. census generation of tabulating machines let population data be sorted by cohort for the first time at national scale. — Image prompt and art direction by Brecht Corbeel; image generated to that direction.

Figure 4. Cohort life tables, assembled decade by decade, gave the UN Population Division its first raw material for projection. — Image prompt and art direction by Brecht Corbeel; generation pending.

Figure 5. Before electronic computing, cohort-component projections were run by hand-cranked adding machines, one age group at a time. — Image prompt and art direction by Brecht Corbeel; image generated to that direction.

Sources

1. United Nations Department of Economic and Social Affairs. About United Nations Population Division. United Nations (2024).
2. Frank W. Notestein. Population: The Long View. Food for the World (University of Chicago Press), ed. Theodore W. Schultz (1945).
3. National Archives and Records Administration. The Great Migration (1910-1970). National Archives (2024).
4. Martin Dribe, Francesco Scalone. The historical fertility transition at the micro level: Southern Sweden 1815-1939. Demographic Research (2014). DOI: 10.4054/DemRes.2014.30.17.
5. E. G. Ravenstein. The Laws of Migration (Second Paper). Journal of the Royal Statistical Society (1889). DOI: 10.1111/j.2397-2335.1889.tb00043.x.
6. United Nations Department of Economic and Social Affairs. 68% of the world population projected to live in urban areas by 2050, says UN. United Nations (2018).
7. Anatoly Vishnevsky. The Demographic Transition in the First World: The Nineteenth Century. Sociological Studies (Sociostudies.org almanac) (2015).
8. Encyclopaedia Britannica editors. Urbanization: Impact of the Industrial Revolution. Encyclopaedia Britannica (2024).
9. Population Index of Population Sciences (PIPS). United Nations Population Division. PIPS / SSDAN (2023).

Originally published at https://absolutedigitalpublishers.com/articles/from-origins-to-frontier-a-history-of-demography-migration-and-urban-futures.