The (Non-)Exceptional City: Sequence, Demography and the Southern Urban Critique

1. Water flowing uphill

In July 1883 Thomas Hardy published an essay on the farm workers of his county in Longman’s Magazine. Much of ‘The Dorsetshire Labourer’ follows the hiring fairs and yearly removals of labouring families between farms, but its sharpest sentence concerns the families who left the villages altogether. ‘This process, which is designated by statisticians as “the tendency of the rural population towards the large towns”,’ Hardy wrote, ‘is really the tendency of water to flow uphill when forced’ (Hardy 1883). The families he had in mind were not, in the first instance, drawn by urban wages. They were the village households outside a farmer’s employ, artisans and small traders among them, who lost their homes as the leases on their cottages fell in and the cottages came down. The statisticians recorded a tendency; Hardy saw a pressure. He returned to the sentence eight years later in Tess of the d’Urbervilles, where the process is ‘humorously designated by statisticians’ and the water is forced ‘by machinery’ (Hardy 1891: ch. 51).

The statisticians’ phrase is now applied to Harare, Lagos and Nairobi far more often than to Dorset. Through the recessions of the 1980s and 1990s the urban populations of many African countries kept growing while urban real incomes fell. Economists had expected falling incomes to slow migration, and for a time the pattern seemed to set African urbanization outside the known relationships between cities and economic growth (Harris 1990; Fay and Opal 2000). Verdicts on the continent have since swung between extremes. A cover of The Economist called Africa ‘the hopeless continent’ in May 2000; another announced ‘Africa rising’ in December 2011 (The Economist 2000; 2011). Through both moods the cities have remained the evidence of last resort for the darker reading. The informal room, the street trade, the overloaded sewer and the stretched municipal budget recur in scholarship and policy as signs that African urbanization is something the world has not seen before.

The London that Hardy’s families reached in the 1880s had seen each of these things. It was a city of rented rooms and casual work, governed until 1855 by some three hundred separate local bodies, whose poor the reading public had just rediscovered in The Bitter Cry of Outcast London (Mearns 1883). Chicago grew more than fiftyfold in the second half of the nineteenth century, faster than Lagos, Kinshasa or Nairobi have grown since 1950. Paris had been rebuilt on borrowed money by a prefect who lost his office in 1870 amid controversy over the borrowing (Pinkney 1957). Each condition now offered as evidence of African exceptionality has a nineteenth-century counterpart, and the counterparts are not obscure. What is unfamiliar is the order in which the conditions arrived, and the demographic engine that drove them. Nineteenth-century cities filled from the countryside while their own death rates kept them from growing from within. African cities have grown, to a large extent, from within.

This paper asks a narrow question of a broad debate: in what respects, if any, is African urbanization historically exceptional? Its answer is that African urbanization is historically familiar in its symptoms and historically distinctive in its sequence, and that the difference is one of lateness rather than of place. African cities did not enter urbanization behind the North, on the same road, and they do not stand ahead of it as a preview of everyone’s future. They entered it later in a shared history: after the medical and public-health technologies that lower mortality had spread far beyond the places that developed them, when no outlet comparable to nineteenth-century transatlantic emigration remained, and as the world’s manufacturing employment was being redistributed. That timing changed the order in which growth, mortality decline, municipal authority and municipal finance arrived. The order, not the symptoms, is where the comparison finds difference.

The argument revises one I have made before. Earlier versions (Muzenda 2019; 2022) answered the claim of African exceptionality by setting nineteenth-century London, Paris and New York beside contemporary African cities, and concluded that cities pass through an ‘urban evolution process’ in which ‘every stage of the evolution has its pains and gains’ (Muzenda 2022: 14). The first half of that argument survives the record; the second does not. A stage model keeps the North as the measure and places African cities in its past, in what Chakrabarty (2000: 8) called the ‘waiting room of history’. The comparison is worth keeping. The ladder is not.

The paper makes three contributions. First, building on Lawhon and Truelove’s (2020) disambiguation of the southern urban critique, it separates the epistemic claim about where urban theory is made, which historical comparison cannot test and does not need to, from the empirical claim that southern cities differ. It then splits the empirical claim in two: a claim of uniqueness, which historical comparison can test and largely refutes, and a claim of configuration, which it can test and partly confirms. Second, it carries into urban theory a finding well established in demography and economic history but little used there: that urban growth in poorer countries since 1950 has been driven to a large extent by urban natural increase, reversing the nineteenth-century pattern (Preston 1979; Dyson 2011; Fox 2012; Jedwab and Vollrath 2019; Menashe-Oren and Bocquier 2021). Third, it replaces the question of whether African cities are exceptional with a question about sequence that can be put to municipal finance and planning practice without placing African cities behind or ahead of anyone.

2. Three exceptionalisms and two displacements

In 1954 Kingsley Davis and Hilda Hertz Golden set the share of national populations living in cities against the share of economically active men working outside agriculture, country by country. Most countries lay close to a common line; a few, Egypt and Korea among them, lay well above it, with more of their people in cities than their industrial employment seemed to warrant. Davis and Golden called the condition ‘over-urbanization’ (Davis and Golden 1954) and did not treat it as necessarily harmful. The term nonetheless carried a benchmark inside it. The discussion that followed set the Asian countries of the 1950s beside Western countries at a comparable level of urbanization. In it, to be over-urbanized was to depart from a relationship between cities and industry that the North was taken to have established, and the departure was read as a disorder of the cities that showed it.

The benchmark was questioned almost at once, and from the South. The Indian economist N.V. Sovani showed that the association between urbanization and industrialization was closer among the poorer countries of 1950 than among the richer. He also showed that the appearance of over-urbanization depended on which Western countries were taken as the norm: set beside Switzerland or Sweden rather than the United States, France, Germany and Canada, the Asian ratios showed no appreciable departure (Sovani 1964; as summarized in Castells 1977: 41–42). Later work in economics turned his doubt into a sustained revision. Fay and Opal (2000) showed that urbanization without growth was, in the words of their title, ‘a not-so-uncommon phenomenon’. Jedwab and Vollrath (2015) set it in historical perspective, and Gollin, Jedwab and Vollrath (2016) distinguished the ‘consumption cities’ of resource-exporting economies, built on rents spent in urban services, from the ‘production cities’ of industrializing ones. Over half a century the economics of urbanization has moved from treating the African and Asian departure from the Northern pattern as a pathology towards treating it as one variant among several. Urban theory has not always followed. The benchmark persists, less as a regression line than as an image of what a city ought to look like at a given level of income.

That image has two inversions in current writing, which the earlier versions of this argument called the apocalyptic and the celebratory. Kaplan’s (1994) ‘The coming anarchy’ read the cities of coastal West Africa as the leading edge of a global breakdown of order, and Davis’s (2006) Planet of Slums generalized the reading into a world-historical claim that informal settlement was the destination of urban growth in the South. Koolhaas and the Harvard Project on the City went to Lagos and returned with what looked like the opposite verdict: ‘Lagos is not catching up with us’, they wrote; rather, ‘we may be catching up with Lagos’ (Harvard Project on the City 2000: 652–653). Comaroff and Comaroff (2012) gave the reversal a programmatic subtitle: Euro-America, they proposed, is evolving toward Africa. Gandy (2005) replied to the Lagos project that its celebration aestheticized a city whose infrastructural crisis had a political history. Watson (2014) has since traced a related optimism into the master-planned new cities promoted across the continent, in which the African city is invited to leap past its own present.

The two readings share a premise that matters more than their disagreement. Each places the African city outside the time of the cities with which it is compared. The apocalyptic reading puts it behind, in the North’s nineteenth century or in a failure to leave it; the celebratory reading puts it ahead, in a future the North has yet to reach. The Lagos text holds both positions at once, describing a city tending towards ‘a terminal condition, a steady state’ and, a few pages on, one that ‘might simply be twenty, fifty or a hundred years ahead of other cities’ (Harvard Project on the City 2000). Fabian (1983) called this move, in anthropology, the denial of coevalness: the placing of the people one studies in a time other than one’s own. Chakrabarty (2000: 7) traced its historicist form, the structure of ‘first in Europe, then elsewhere’, in which other societies wait for a history that Europe has already had. Neither wrote about cities, but the displacement they describe organizes much of the writing on African urbanization, whatever its tense.

Ferguson (2006) observed a further turn. As the developmentalist narrative lost credibility in the 1980s and 1990s, Africa’s position in the world was read less as a stage on the way to modernity and more as a fixed status in a global hierarchy: a difference of place rather than of time. Exceptionalism in its strongest form performs that conversion. The African city becomes a different kind of city, rather than an earlier or a later one. Mbembe’s (2001) account of postcolonial time as an entanglement of pasts, presents and futures, rather than a succession of stages, and Simone and Pieterse’s (2017) account of urban life as the inhabiting of ‘dissonant times’, offer a way out of both the temporal and the spatial versions. Both insist on coevalness without denying that the African present is composed differently, and it is in that spirit that this paper approaches the record.

Against this background the claim that African, or more broadly southern, cities are exceptional needs to be taken apart before it can be tested. Lawhon and Truelove (2020) distinguish three versions of the southern urban critique: that cities of the global South differ empirically from those on which urban theory was built; that Euro-American intellectual dominance has marginalized other traditions of urban thought; and that postcolonial engagement requires attention to the practices of research itself. Historical comparison bears only on the first. The question with which the African Centre for Cities’ International Urban Conference closed in Cape Town in February 2018, whether the field needed one theory of southern urbanism or many (Muzenda 2022: 13), assumed that the southern claim was one claim. It is several, and only one of them is open to historical test. Within that empirical claim, two further claims need to be held apart. A claim of uniqueness holds that southern cities exhibit conditions absent from the urban past of the North: growth without industrialization, mass informality, ungovernable fragmentation. A claim of configuration holds that the conditions may each be familiar but combine distinctively. The first can be refuted by counter-example. The second cannot; it can be tested only by comparing combinations and orders of conditions across cases.

The leading southern theorists have mostly made the epistemic and configurational claims, not the claim of uniqueness. Robinson’s (2006) case for ‘ordinary cities’ was explicitly a case against the developmentalist hierarchies that sorted cities into the modern and the not-yet-modern; it is, in this paper’s terms, an argument against exceptionalism of any kind. Roy (2011) welcomed ‘subaltern urbanism’ as a departure from apocalyptic accounts of the slum but questioned its celebration of entrepreneurialism and self-organizing economies, and proposed concepts, from peripheries to ‘gray spaces’, with which to rethink the urban as a whole. Schindler (2017) proposed three tendencies, a persistent disconnect between capital and labour, discontinuous and contested metabolic configurations, and the co-constitution of political economy with urban materiality, as the elements of a paradigm of southern urbanism: a configurational claim. Bhan (2019) has carried the critique from theory into practice, proposing squatting, repair and consolidation as southern urban practices and calling for a broader vocabulary built from them. From the other side of the debate, Scott and Storper (2015) and Storper and Scott (2016) defend a general theory of cities grounded in agglomeration and the urban land nexus, and treat southern difference as variation within it. The earlier versions of this argument read the southern field as a single camp asserting uniqueness. Much of what they rejected was not there to reject; what was there, the configurational claim, they did not test.

One further term is needed, because every version of the comparison, the earlier ones included, measures African cities against something. The measure is rarely the nineteenth-century city as its residents knew it. It is the North as remembered after its urban transition was complete: sewered, policed, consolidated under a single municipal authority, employed in factories and offices, and housed by owner-occupation. Chakrabarty’s (1992; 2000) ‘hyperreal Europe’ names the silent referent of non-Western history in general. What this paper calls the remembered North is its urban and retrospective form: not Europe as such, but the industrial city as it appeared once its transition had run its course, read back into the decades when the transition was still under way. Its effect is to make the outcome of a long and contested process look like a starting condition, and so to make any city still inside the process look deficient. Section 4 sets the remembered North against the archive.

The paper therefore does three things with the exceptionality claim. It accepts the epistemic critique, which historical comparison cannot touch. It tests the uniqueness claim against the nineteenth-century record and finds it largely false. And it restates the configurational claim as a claim about sequence, which the record supports. The restatement matters because a sequence can be compared across many cases, South–South as well as North–South, without making any one of them the norm. It also joins a growing attention to time in southern urban theory, from Mbembe’s (2001) account of entangled temporalities to Simone and Pieterse’s (2017) ‘dissonant times’.

3. Comparing sequences, not snapshots

In 1979 Samuel Preston asked a simple question of the United Nations’ urban data: where did the growth of cities in developing countries come from? Urban populations grow in three ways: through the excess of births over deaths among urban residents (natural increase), through net migration from the countryside, and through the reclassification of settlements from rural to urban. Preston’s reappraisal found that in the 1950s and 1960s most urban growth in developing countries came from natural increase rather than migration (Preston 1979). Later estimates have varied widely by country and decade (United Nations 2001), but the pattern has held, and direct census measurement has since confirmed natural increase as the main contributor in Africa (Bocquier, Menashe-Oren and Nie 2023). The finding changes what a comparison with the nineteenth century has to compare. A snapshot of two cities at similar levels of crowding or informality says nothing about how either came to be that way. A comparison of sequences does.

Four terms are used in exact senses throughout. Urban growth is the rate of change of an urban population, expressed here as an average annual exponential rate. urbanization is the rate of change in the urban share of a national population; a country can combine rapid urban growth with slow urbanization if its rural population is also growing fast. Informality refers to work, building or settlement that takes place outside the rules a state applies to it. It is not a synonym for poverty, and much informal activity is neither poor nor marginal (Roy 2005). Municipal authority refers to the local government of a city, as distinct from the national state and from the fiscus, the national treasury on which municipal borrowing has historically depended. The earlier versions of this argument used several of these terms interchangeably, most consequentially ‘urbanization rate’ for what was in fact urban growth.

The comparison distinguishes symptoms from sequence. A symptom is an observable condition of a city at a given time: casual labour, rented rooms, unsewered districts, overlapping authorities, municipal deficits, the alarm of observers. A sequence is the order of four transitions: in the demographic engine of urban growth, between in-migration and natural increase; in mortality, between an urban penalty and an urban advantage; in municipal authority, between fragmentation and consolidation; and in municipal finance, between incapacity and sustained borrowing for infrastructure. Two conditions external to the city frame these transitions: whether an outlet existed for surplus labour through emigration, and how much industrial employment the world economy offered at the time of the transition. The paper’s claim is that African and nineteenth-century Northern cities share their symptoms and differ in their sequence.

The nineteenth-century cases are London, Paris, New York and Chicago between roughly 1830 and 1900. They are chosen not as norms but because they are the referents the exceptionality debate invokes, explicitly in the over-urbanization literature and implicitly wherever the remembered North sets the standard. Huchzermeyer’s (2011) comparison of nineteenth-century Berlin with twenty-first-century Nairobi supplies a fifth case and is this paper’s closest methodological ancestor: a Johannesburg-based scholar showing that a housing form treated as novel in Nairobi had a precise precedent in a European capital, and that the precedent’s regulatory history held lessons for the present. The design follows Robinson’s (2016; 2022) case for comparative tactics that can begin from any city and need not presume a norm. The African cases are Lagos, Kinshasa, Nairobi and Harare since 1950. They span West, Central, East and Southern Africa, include cities with and without a history of settler rule, and cover a range of sizes and growth paths. Latin American cities between 1940 and 1980 serve as a South–South control. They too grew rapidly after mortality had fallen, but with more of their labour absorbed by import-substituting industry (Gilbert and Gugler 1992), and their residents were more integrated into urban labour markets than the image of marginality allowed (Perlman 1976). The control helps separate the effects of demographic timing from those of labour demand. The design has an asymmetry that should be stated. The nineteenth-century sequence is documented here at city level, chiefly for London, whereas the African sequence is documented at the level of the regional urban transition, with the four cities serving as illustrations rather than fully reconstructed cases. Reconstructing each African city’s sequence is the next step the argument requires.

The comparison rests on a proposition about timing with a long lineage outside urban studies. Gerschenkron (1962) argued, against stage theories of growth such as Rostow’s (1960), that countries industrializing late did not repeat the path of the pioneers: they found substitutes for the prerequisites the pioneers had needed, and late arrival changed the route itself. Pierson (2004) generalized the point for the social sciences: when something happens shapes how it happens, and the same elements in a different order produce different outcomes. Applied to cities, the proposition is that a city’s position in world time, not its location on a single developmental track, conditions the sequence it can follow. Jedwab and Vollrath (2019) have in effect shown this for the demographic engine of urban growth, and Goodfellow (2022) has used the frame of late urbanization to explain why three East African cities diverged. This paper carries the same logic into the exceptionality debate.

The evidence comes from standard historical scholarship on the nineteenth-century cases; from primary texts where they are accessible, among them Hardy, Goschen and White; from the United Nations’ estimates of city populations (UN DESA 2019); and from demographic decompositions of urban growth. African urban data carry known weaknesses. The Democratic Republic of Congo has completed only one general census, in 1984, so Kinshasa’s population is a projection. National definitions of ‘urban’ vary so widely that a harmonized global definition finds Africa’s urban share considerably higher than national statistics suggest (Dijkstra et al. 2021). The paper therefore relies on precise figures only where the argument turns on orders of magnitude that survive these uncertainties.

The argument can be wrong in specific ways. If nineteenth-century Northern cities were found to have grown mainly by natural increase, or without access to emigration, or with urban mortality falling before municipal sanitary investment, the distinctiveness of the African sequence would collapse. If contemporary African cities were found to grow mainly by net in-migration under an urban mortality penalty, it would collapse from the other side. And if the sequence proved specific to Africa, absent from other regions that urbanized late, the argument from lateness would give way to an argument from place.

A word on position. I write as a Zimbabwean town planner practising in South Africa, and as the author of the polemical earlier versions that this paper revises. The practitioner’s instinct is to judge: to read an overflowing sewer as a failure and a by-law as a cure. The method is meant to discipline that instinct by describing what each arrangement did, and in what order, before asking what follows. Where the practitioner’s knowledge helps is in knowing how municipal instruments actually bind, which is where Section 7 ends.

4. Familiar symptoms

The phrase that opens this paper was written about Britain. When Geddes (1915: 128) wrote ‘Slum, Semi-slum, Super-slum’, he was describing the industrial towns of his own country, and the phrase has since migrated, with the word ‘slum’ itself, to the cities of the South. This section follows five conditions that the exceptionality claim treats as distinctively African back into the nineteenth-century record. None is the whole picture; each is enough to show that the condition, taken alone, is not unique.

Fragmented authority

In 1855 The Times reported that the local administration of London was ‘carried on by no fewer than 300 different bodies’ (quoted in House of Commons Library 1997). They included parish vestries and the boards and trusts for paving, lighting, sewers and improvements on which some ten thousand commissioners sat, each body with its own boundaries, rates and statutory powers (Sheppard 1971). None was a planning authority in any modern sense, and the City of London, the one body with full corporate standing, had been left outside the municipal reform of 1835. The Metropolis Management Act of 1855 placed a Metropolitan Board of Works above the vestries, but fragmentation was a national condition, not a metropolitan one. Introducing the reform of local taxation in 1871, George Goschen told the House of Commons that ‘we have a chaos as regards authorities, a chaos as regards rates, and a worse chaos than all as regards areas’ (HC Deb 3 April 1871, vol. 205, col. 1116). Twelve years later a standard textbook still described English local government as ‘a chaos of areas, a chaos of authorities, and a chaos of rates’ (Chalmers 1883: 17).

Cirolia’s (2020) account of ‘fractured fiscal authority’ in African cities describes a structurally similar condition. Revenue powers, spending responsibilities and infrastructure mandates are divided between national ministries, parastatal utilities, regional governments and municipalities, so that no single authority can plan, finance and maintain a network as a whole. Lagos is governed by a state government, twenty local government areas and federal agencies with overlapping mandates; Kinshasa is divided into twenty-four communes; the built-up area of Harare spans several separate local authorities, Chitungwiza and Epworth among them. The vocabulary has changed, from chaos to fracture, but the condition is recognizably the one Goschen described.

Rented rooms

The urban poor of the nineteenth-century North rented, and the poorest rented by the room or the bed. In England three-quarters of all households rented privately as late as 1918, and owner-occupation became the majority tenure only around 1971 (DCLG 2015). Common lodging houses, where a bed could be had by the night, were brought under police registration by an Act of 1851. In Paris the garnis, the furnished lodging houses, sheltered much of the migrant population of the July Monarchy, and Chevalier (1973: 217–218) describes newcomers who ‘settled in inferior occupations and inferior districts on the margin of the capital’. In New York the tenement became the characteristic working-class dwelling, first regulated by the Tenement House Act of 1867 and more stringently in 1901 after a decade of agitation that Riis’s (1890) How the Other Half Lives had helped to start. Berlin’s Mietskasernen, the ‘rental barracks’ of the late nineteenth century, housed much of the city’s growing population in courtyard blocks of single and double rooms (Huchzermeyer 2011).

Huchzermeyer (2007; 2011) found the same form in Nairobi: multi-storey tenement districts built from the 1990s by private landlords, in which most residents rent single rooms, and whose closest historical analogue is the Berlin Mietskaserne rather than anything in the planning literature on African informal settlement. In Harare, lodging in rooms let by householders in the high-density suburbs has long been a principal source of shelter for the city’s poor (Rakodi 1995). Housing policy built on the remembered North has nonetheless treated home-ownership as the norm and renting as a transitional or deviant tenure. The nineteenth-century record suggests the reverse: the rented room was the ordinary tenure of urban growth, and mass owner-occupation a much later outcome.

The street economy

Henry Mayhew’s London Labour and the London Poor opens with the street-folk: costermongers, street-sellers, street-finders, performers and labourers, whom he counted in the tens of thousands (Mayhew 1861–62). Stedman Jones (1971) later showed how much of late-Victorian London’s labour market remained casual, with dock work hired by the half-day, seasonal trades and sweated work at home, and how the casual poor were concentrated, chronically underemployed, in the inner districts. In Paris, the ‘dangerous classes’ of Frégier’s (1840) treatise were in large part workers without regular employment.

The concept now used to describe such work was coined not in London or Paris but in Accra. Keith Hart’s study of ‘informal income opportunities’ among migrants in the Accra suburb of Nima, presented in 1971 and published in 1973, and the International Labour Office’s employment mission to Kenya in 1972, gave urban studies the vocabulary of informality (Hart 1973; ILO 1972). The categories through which the street economy of the nineteenth-century North is now read were made in African cities and only then turned back on the archive. That is theory from the South in the most practical sense, and it required no claim of exceptionality to do its work.

Clearance and its oscillations

Clearance was an instrument of nineteenth-century urban government before it was one of twentieth-century planning. Between 1853 and 1870 the Second Empire’s works cut boulevards through the densest quarters of central Paris and moved much of their working population towards the periphery (Pinkney 1958; Harvey 2003). In London, New Oxford Street was driven through the St Giles rookery in the 1840s; railway building and street improvements demolished working-class housing over the following decades; and the Artisans’ and Labourers’ Dwellings Improvement Act of 1875 gave the clearance of unhealthy areas a statutory form, under which displacement repeatedly outran rehousing (Wohl 1977; Yelling 1986). In New York, Mulberry Bend, the most notorious block of the Five Points district, was demolished in the 1890s and replaced by a park (Anbinder 2001).

The instrument recurs in African cities. In 1990 the Lagos State government cleared the Maroko settlement on the city’s coastal fringe (Agbola and Jinadu 1997). Between May and July 2005, the Government of Zimbabwe’s Operation Murambatsvina, officially ‘Operation Restore Order’, demolished structures it classified as illegal in Harare and other towns, and the United Nations Special Envoy estimated that some 700,000 people lost their homes, their livelihoods or both (Tibaijuka 2005). Demolitions of informal trading structures in Harare recurred during the COVID-19 restrictions of 2020 (Mwonzora 2022). Kamete (2009) examined the part played by planning law and planners in the 2005 operation. His later work offers an analytical rather than a normative reading of the recurrence. Drawing on Bauman’s figure of the stranger, Kamete (2020) argues that informality presents itself to planners as ‘neither friend nor enemy’, and that this unresolved ambivalence, rather than settled hostility, drives its invalidation.

The nineteenth-century record shows a comparable oscillation between toleration, regulation and clearance, often within a single decade (Wohl 1977; Yelling 1986), as authorities without the means to rehouse acted on districts beyond their administrative reach. It also complicates one configurational claim directly. Schindler (2017: 53) contrasts a nineteenth-century North, where turning migrants into disciplined workers was the raison d’être of urban government, with a South in which governments transform territory rather than improve populations. Second Empire Paris was a programme of territorial transformation, financed by borrowing against the expected rise in land values (Pinkney 1957; Harvey 2003), and it did comparatively little to house the migrants it displaced. The contrast Schindler draws is real, but it is a difference of degree and of moment rather than of kind.

The alarm of observers

The last symptom is the observer’s alarm. The Bitter Cry of Outcast London (Mearns 1883) presented the housing of the London poor as a moral emergency; Riis (1890) did the same for the tenements of the Lower East Side; Frégier (1840) gave a durable name to the fear that the urban poor of Paris threatened public order. Kaplan’s (1994) account of West African cities belongs to the same genre, as the earlier version of this argument observed (Muzenda 2022: 3). What differs is the tense. Nineteenth-century alarm described a present and demanded action within it. Much writing on African cities projects a future, in which the number of urban residents expected in thirty or fifty years is itself the emergency, although projections of African urban growth have repeatedly been revised and their base data are often thin (Potts 2009; 2012).

The remembered North, dissolved

Set side by side, these records dissolve the remembered North. The yardstick city, sewered, consolidated, owner-occupied and formally employed, is the outcome of a transition that ran through the second half of the nineteenth century and, in housing tenure, deep into the twentieth. Its starting condition looked much like the African city of exceptionalist accounts. Contemporary verdicts of failure proved provisional too. In 1890 Andrew D. White, the first president of Cornell University, judged ‘the city governments of the United States’ to be ‘the worst in Christendom—the most expensive, the most inefficient, and the most corrupt’ (White 1890: 357). Nearly a century later Teaford (1984) re-examined the same governments over the years 1870 to 1900 and found an ‘unheralded triumph’: water supplies, street lighting, parks, libraries and transport systems built at a scale that, by 1900, compared with any in the world, while the reputations of those governments were being written by their critics.

The revision cuts both ways, and the ambivalence is the finding. Teaford’s reading warns against treating the present condition of African municipal government as permanent; contemporary verdicts on nineteenth-century cities were wrong in the same way. It does not promise the same outcome, because the American cities’ achievement rested on a sequence of demographic, fiscal and institutional changes that Section 5 shows African cities have not shared. The symptoms match. Whether the resolution can match depends on the order in which they arrived.

5. A distinctive sequence

Rate

Chicago had 29,963 residents in 1850 and 1,698,575 in 1900. Its population doubled, on average, every nine years for half a century, a pace that none of the African cities in Table 1 has sustained over a comparable period. Table 1 sets the nineteenth-century cases beside the African ones.

The table does not support a simple contrast between slow Northern and fast African growth. London and Paris grew at about 1.6 per cent a year across the century, doubling every four decades; Lagos and Kinshasa have grown three to four times as fast. The American cities of the late nineteenth century, however, grew faster still: Chicago at 8 per cent a year. The earlier versions of this argument were right that rapid urban growth is not unique to Africa and wrong to rest the case on London and Paris, whose own figures showed a difference of rate. Rate distinguishes some African cities from some nineteenth-century cities. It does not distinguish African urbanization from nineteenth-century urbanization. Nor is rate the clearest measure of what is distinctive. Africa remained mostly rural in 2018, with 43 per cent of its population living in urban areas (UN DESA 2019). Its cities have grown quickly while urbanization, the rise in the urban share, has been comparatively slow, because its rural population has continued to grow as well. If there is a difference, it lies in what has driven the growth.

Engine

Nineteenth-century cities grew by drawing people in. Their death rates were higher than those of the surrounding countryside, a condition historical demographers call the urban penalty (Kearns 1988; Woods 2003). New estimates of life expectancy at birth show Britain’s largest provincial cities losing ground during the 1830s and 1840s, with expectation of life well below the national level (Szreter and Mooney 1998). Through most of the century, therefore, the growth of such cities depended on a continuous stream of migrants, which is why contemporaries, Hardy among them, wrote about urbanization as a question of migration. Chicago and New York grew above all through migration from Europe and from the American countryside. Natural increase became a major source of urban growth in Britain only as urban mortality fell from the 1870s, after the sanitary investment of the preceding decades (Woods 2003; Chapman 2019).

African cities since 1950 have grown differently. Mortality fell across the continent from the 1950s, and fell faster in cities than in the countryside, so that urban natural increase became a principal source of urban growth from early in the transition (Jedwab and Vollrath 2019). Jedwab and Vollrath’s calibration suggests that between 1950 and 2005 the decline in urban mortality could have doubled both the urbanization rate and the size of informal urban areas in their sample of poor countries. Menashe-Oren and Bocquier (2021), decomposing growth in low- and middle-income countries between 1985 and 2015, find that the main push to urbanization came from differences in natural increase between urban and rural areas, owing mostly to lower under-five mortality in cities, and that the role of migration and reclassification was small. Direct census estimates for Africa point the same way (Bocquier, Menashe-Oren and Nie 2023). Fertility decline has meanwhile been slower in sub-Saharan Africa than in other regions at comparable stages of the transition (Bongaarts and Casterline 2013), which sustains urban natural increase.

The two engines produce cities that look alike and behave differently. A city that grows by in-migration can, in principle, slow its growth by changing the conditions that draw migrants or push them from the land, and nineteenth-century observers argued over exactly those levers. A city that grows from within cannot, because most of its growth is born to people already resident. Policies premised on migration as the main source of urban growth address a diminishing share of that growth in African cities. The earlier versions of this argument sensed the importance of demography but located it in the wrong century, claiming that falling death rates had driven the growth of Paris and the English cities in the first half of the nineteenth century (Muzenda 2022: 8). The claim was false of the nineteenth-century North; it is true of African cities now.

Outlets

Nineteenth-century cities also had a way out. About 55 million Europeans migrated to the New World between 1850 and 1914 (Hatton and Williamson 1998). Emigration drained labour from the European countryside and from European cities, and contributed to the convergence of real wages between the Old World and the New. The cities of nineteenth-century America were in part the destination of that outflow: Chicago’s growth was Europe’s emigration seen from the receiving end. African migration, by contrast, remains overwhelmingly intra-continental (Flahaux and de Haas 2016). No outlet on the scale of the transatlantic migration has been open to the African urban transition, which has had to absorb its own growth.

Order

In London the order ran from fragmentation through consolidation to infrastructure, and only then to lower mortality. The Metropolitan Board of Works was created in 1855. The ‘Great Stink’ of the summer of 1858 moved Parliament, within weeks, to authorize the Board to borrow £3 million for main drainage under a national guarantee (HC Deb 19 July 1858). Bazalgette’s intercepting sewers were built through the 1860s and 1870s (Sheppard 1971). Mortality in British cities fell markedly only from the 1870s, and recent estimates attribute an important share of the decline to investment in water and sewerage (Chapman 2019; see also Szreter 1988). In major American cities, clean water accounted for nearly half of the total reduction in mortality in the early twentieth century, three-quarters of the reduction in infant mortality and nearly two-thirds of the reduction in child mortality (Cutler and Miller 2005). The investment itself came late: Williamson (1990) argues that Britain under-invested in urban social overhead through the early decades of industrialization. Lagging infrastructure finance is therefore a familiar symptom too. What the nineteenth-century record adds is its place in the order: the decline in urban mortality was produced, in substantial part, by municipal capital works financed by municipal borrowing, and it followed them.

In African cities the order has been reversed. Urban mortality fell from the 1950s through vaccination, antibiotics and, later, oral rehydration, technologies developed elsewhere and diffused internationally, before the water and sewerage networks of most cities had reached the populations they served (Jedwab and Vollrath 2019). Mortality decline did not have to wait for the sewer. The sanitary deficit that had made nineteenth-century cities lethal, and so kept them from growing from within, no longer acted as a brake on growth; it persisted as a hazard instead. The largely urban cholera epidemic that Zimbabwe experienced from mid-2008, with more than 92,000 cases and 4,000 deaths by early 2009 (Mason 2009), recalls London’s epidemics of 1832, 1849, 1854 and 1866. It struck cities whose growth had not waited for their pipes.

The fiscal side of the sequence differs in the same way. London’s Board could borrow £3 million in 1858 because a national treasury stood behind it and a deep domestic capital market would lend. Second Empire Paris borrowed through the Caisse des Travaux de Paris and through bonds placed with the Crédit Foncier (Pinkney 1957). Only a handful of sub-Saharan cities have borrowed on domestic capital markets, and the legal and regulatory conditions for municipal borrowing remain the binding constraint in most (Paulais 2012; Gorelick 2018). Revenue powers are themselves divided (Cirolia 2020), and much large urban infrastructure has been financed by national governments and external lenders filling gaps that municipal finance cannot (Goodfellow 2020). The instrument that resolved the nineteenth-century sequence, consolidated municipal borrowing for sanitary capital, is the one least available to the cities whose growth now most requires it.

Labour

The last element is labour. Nineteenth-century urban labour markets were casual at the bottom, as Section 4 showed, but industry absorbed a large and rising share of urban workers in Britain and later in the United States, Germany and France (Bairoch 1988). African cities have urbanized with less industrial employment. Gollin, Jedwab and Vollrath (2016) show that in resource-exporting economies urbanization has concentrated in consumption cities whose employment lies mainly in non-tradable services. Rodrik (2016) finds that manufacturing’s share of employment now peaks earlier, at lower levels and lower incomes, in countries that industrialize late. Latin American cities in the decades after 1940, the paper’s South–South control, absorbed more of their growth in import-substituting industry (Gilbert and Gugler 1992). African cities, urbanizing later, met a world economy in which that route had narrowed.

Lateness rather than place

Taken together, these elements describe a sequence rather than a list of symptoms. In the nineteenth-century North, migration drove urban growth under an urban penalty; emigration relieved pressure on labour markets; municipal authority was consolidated and given borrowing powers; sanitary capital was built; urban mortality fell; and only then did cities grow substantially from within, by which time industry was absorbing much of their labour. In African cities since 1950, mortality fell first, through technologies that required no municipal capital; cities grew from within before authority was consolidated or borrowing powers secured; no outlet comparable to transatlantic emigration existed; and manufacturing offered a narrower route into urban employment. The elements are the same. The order is reversed at the two points that matter most: mortality decline now precedes sanitary investment, and growth from within precedes the fiscal instruments needed to serve it.

I call this a difference of lateness rather than of place, and the term’s ancestry should be stated. It extends Gerschenkron’s (1962) argument that latecomers do not repeat the pioneers’ path but meet a world the pioneers have changed, and it follows Goodfellow’s (2022) use of late urbanization for East Africa. It also stands on demographic accounts of urbanization as a global historical process. Fox (2012) traces Africa’s urban transition to innovations in disease control that originated in Europe and spread through colonialism, trade and development assistance, and Dyson (2011) places urbanization within the demographic transition itself. What lateness adds is the order in which those innovations arrived relative to municipal authority and finance, and what that order means for the exceptionality debate. It differs from ‘behind’ because it insists on coevalness (Fabian 1983): African cities share the North’s present and its world economy, and their sequence differs because the pioneers’ own transitions altered the medical, migratory and industrial conditions under which later transitions occur. It differs from ‘place’ because it refuses an essential African condition (Ferguson 2006): the latecomer sequence is visible, to varying degrees, wherever urbanization began after the global diffusion of mortality-reducing technology, in Latin America and Asia as well as Africa (Preston 1979; Jedwab and Vollrath 2019). Africa’s sequence is the most pronounced instance of a general pattern, not an exception to one.

The concept has a boundary. Lateness does not mean that Africa came to cities late; Section 6 shows that it did not. It refers to the modern urban transition, the passage towards majority-urban societies under industrial-era demography, and it holds only as far as that transition does.

6. Before and beneath the sequence

In 1929 Gertrude Caton-Thompson excavated at Great Zimbabwe for the British Association for the Advancement of Science, which met in South Africa that year. Her report confirmed what David Randall-MacIver had concluded after the Association’s previous visit in 1905: that the stone enclosures south-east of present-day Masvingo were built by the ancestors of the region’s African population, between roughly the eleventh and the fifteenth centuries (Randall-MacIver 1906; Caton-Thompson 1931). Some settler-era writing disputed the conclusion for decades afterwards, and Chirikure (2021) has described the long work of reclaiming what he calls a ‘confiscated’ past. The site was the capital of a state whose centre housed thousands of people and whose trade reached the Indian Ocean coast.

Great Zimbabwe is one instance of a long urban history. At Jenne-jeno on the middle Niger, excavation has traced a settlement occupied from the last centuries BCE that had become a large urban cluster by the end of the first millennium CE, without the citadel or palace that urban theory once took as the mark of a city. McIntosh and McIntosh (1993) called such places ‘cities without citadels’, and R.J. McIntosh (2005) has read them as self-organizing urban landscapes in which authority was distributed rather than concentrated. Aksum, Ife, Benin City and the stone towns of the Swahili coast, Kilwa among them, belong to the same record (Connah 2001; Coquery-Vidrovitch 2005). The idea of the city that comparative urbanism inherited, built from the walled and hierarchical towns of Mesopotamia and the Mediterranean, had difficulty recognizing some of these places as cities at all.

The record matters to the argument in two ways. It sets a boundary on lateness: Africa did not come late to cities, only to the modern transition towards majority-urban societies under industrial-era demography. And it corrects the remembered North from another direction. The over-urbanization benchmark and its successors measured African cities against a Northern urban history that began, for comparative purposes, with industrialization. The longer African record makes the industrial city one urban form among several, and a recent one.

In parts of the continent the modern transition was regulated by law in ways the nineteenth-century North matched only in part. English settlement law allowed parishes to remove paupers to their parish of legal settlement (Snell 2006), but it was not defined by race and did not tie residence in towns to employment. South Africa’s Natives (Urban Areas) Act of 1923 and Southern Rhodesia’s Land Apportionment Act of 1930 treated African urban residence as conditional and temporary, confined it to designated locations and tied it to employment (Raftopoulos and Yoshikuni 1999; Yoshikuni 2007). Salisbury, now Harare, was planned on the premise that its African residents were workers whose families belonged in the countryside (Yoshikuni 2007). Potts (2010) has shown how the circular migration between town and country formed under these controls persisted long after them, and how it continues to shape the size and composition of Zimbabwe’s urban population.

The controls shaped the composition of urban growth at least as much as its rate. On United Nations estimates Harare grew by about 5.5 per cent a year in the 1950s and 5.2 per cent in the 1960s, slowed to 3.9 per cent in the 1970s, and rose again to 5.3 per cent in the 1980s, the first decade after independence, as restrictions on African urban residence fell away (UN DESA 2019). Part of the post-1980 growth was the arrival of households that colonial urban policy had kept in the countryside (Potts 2010). Since 2000 the city’s estimated growth has been far slower, although estimates for these decades depend heavily on where its boundary is drawn. On this evidence, colonial urban administration in the settler territories did less to delay urbanization than to determine who could reside in towns, and on what terms. The earlier versions of this argument held that urbanization in Africa ‘was delayed by colonial restrictions’ (Muzenda 2022: 14). The claim needs a boundary. It applies most clearly to the settler colonies of eastern and southern Africa, and much less to the long-urbanized societies of West Africa, where colonial Lagos grew as a port on an older Yoruba and Bini urban landscape.

The comparison therefore stops at four points. It cannot test the epistemic claim about where urban theory is made, and does not try to. It cannot carry the legal regulation of urban residence by race under colonial rule, which shaped African urban growth in ways the London and Paris cases cannot illuminate. It cannot carry the scale of the largest African cities: Lagos and Kinshasa, each estimated at more than 14 million people in 2020 (UN DESA 2019), are larger than any city of the nineteenth century. And it cannot carry the climate risks to which African cities are now exposed, from coastal and riverine flooding to extreme heat, which no nineteenth-century city faced in that form (Trisos et al. 2022). The last three are the places where African Urbanization is distinctive, and where the word ‘exceptional’ still has work to do. Even there, the distinctiveness is historical and conjunctural, a product of when the transition occurred and under what regime, rather than a property of African cities as such.

7. Conclusion: a sequence question

The record answers the question this paper began with, though not in the form in which the debate has usually asked it. On symptoms, the nineteenth-century archive is plain. The rented room, the casual street economy, fragmented authority, lagging infrastructure finance, clearance and the alarm of observers were all present in London, Paris, New York and Chicago, and some, such as the tenement, had precise formal analogues in African cities. On rate, the archive is mixed: London and Paris grew far more slowly than Lagos or Kinshasa, and Chicago far faster. On sequence, it is plain again, in the other direction. The engine of growth, the outlet for labour and the order in which mortality, authority and finance changed all differ, and they differ because of when the African transition began.

For southern urban theory this has three implications. The epistemic critique stands untouched: nothing in a historical comparison can show that theory built from a few Northern cities travels well. The claim of uniqueness does not survive the archive and is better set aside, not least because its strongest form converts a difference of time into one of place (Ferguson 2006). The claim of configuration, restated as a claim about sequence, becomes testable and portable. It can be put to Latin American and Asian cases, and it reads back onto the North. Seen from Lagos, the Northern sequence looks less like a norm than like the particular route of a pioneer, which had to build its own decline in urban mortality out of municipal capital because the means to deliver it otherwise did not yet exist. That is southern urban theory in the spirit of Robinson’s (2006) call for an urban studies built from a world of ordinary cities, and it needs no claim that African cities are exceptional.

For practice, the remembered North has been a poor guide because it recommends an endpoint, the sewered, consolidated, owner-occupied city, without the sequence that produced it. Instruments taken from that endpoint, from building standards to the clearance of unhealthy areas, assume a city that grows by migration and has the fiscal means to rehouse; African cities grow largely from within and mostly lack those means. The Victorian sequence offers a narrower and more useful lesson. Municipal authority was consolidated; it was given the power to borrow under a national guarantee; and the borrowing was spent on sanitary capital before, and as a condition of, the decline in urban mortality. In African cities mortality fell first, and the municipal instruments have yet to be assembled. The practical question is therefore no longer how to slow urban growth, most of which is now born in the city, but which instrument has to come first when growth arrives before the means to serve it.

Hardy’s water was forced uphill from the countryside, by the pulling down of cottages that contemporaries could see and argue about. In Lagos, Kinshasa and Harare much of it now rises from springs within the city. The nineteenth-century city outgrew its crisis through a sequence of consolidated authority, municipal borrowing, sanitary investment and the outlet of emigration that African cities cannot replay in that order. What remains is the question the Victorian record cannot answer for them: when a city’s growth is generated from within, before its fiscal instruments exist, which instrument must come first, and which tier of government is in a position to build it?