The global future will be shaped by urbanisation - a reality that has focused much scientific attention on so-called ‘global cities’. This preoccupation with the metropolises of the global North contrasts starkly with the unfolding dynamics of urban transition and risks marginalising the poor as the world’s population becomes increasingly urbanised. In demographic terms, the bearers of future urbanisation trends will not be a well-connected class of global consumers, but the mass of rural residents found mainly in the countries of the global South. Moreover, much of this urbanisation is likely to involve small towns, rather than large cities. While the empirical realities of prospective urbanisation patterns will differ from the historical experiences of the global North, the theoretical tools available to analyse the distributional consequences of this process have nonetheless been developed mainly in the context of the global North. Marxist-inspired perspectives stress the inherent inequality resulting from urbanisation as global capital is increasingly concentrated in cities, resulting in spatial, social and economic differentiation (Theodore et al., 2011). Such perspectives can be contrasted with models of structural transformation and economic agglomeration which instead hypothesise a positive link between urbanisation, economic growth and poverty reduction as societies become increasingly urbanised and economies grow (Ravallion et al., 2007; Timmer, 2009).
Evidence from developing countries suggests that theories of urban transition modelled on the historical experience of the global North have been outrun by the unravelling dynamics of urban change, however (Rigg et al., 2016). One explanation for the poor fit between standard models of urban transition and the unfolding processes of urbanisation is found at the micro level, where multi-sited, oscillatory livelihoods challenge historical accounts that involved a cleaner break between urban and rural livelihoods. Another crucial demographic difference concerns the gendered composition of current migration flows, with studies suggesting that they are increasingly being driven by young women, rather than historically based patterns of male selectivity (Montgomery et al., 2016). In sub-Saharan Africa specifically, urbanisation is generally depicted as distress-driven (Andersson Djurfeldt, 2015).
Geographers working in the context of the global South have recently challenged the relevance of neo-Marxist perspectives for understanding urbanism in the global South where the logics of capital accumulation offer only a partial explanation for urbanisation (Robinson, 2011; Parnell and Pieterse, 2016). These calls for comparative urbanism are, however, largely concerned with understanding spatial dynamics in particular cities, rather than the broader process of urbanisation per se.
The poor theoretical understanding of the specific type of urbanisation of most relevance to the urban transition - that is the growth of small towns - aggravates a general analytical and empirical mis-match. The distributional aspects of lower level urbanisation are undertheorised in the sense that we lack analytical models that place the dynamics of localised urban growth in relation to broader urban systems, rural surroundings and the livelihoods and social relations of people living in small towns (towns with up to 50,000 inhabitants). In turn this gap is related to the monoscalar approach that characterises current theoretical perspectives. This contribution departs from existing theoretical perspectives and recent empirical literature and proposes a multi-scalar, nested approach to analysing small town growth in the context of sub-Saharan Africa specifically. Several empirical trends point to the need for rethinking and adapting theoretical perspectives on lower level urban growth to African realities. Not doing so jeopardises the prospects for the social and economic inclusion of the poor in urbanisation processes.
Urbanisation trends and structural transformation theory
At 41.4 per cent, sub-Saharan Africa has the lowest share of urban population in the world, a share which is expected to grow rapidly over the coming decades. Not only will sub-Saharan Africa be at the forefront of the urban transition, but it will also be the region that increasingly drives global population increases, while containing a large share of the global poor. Historically, urbanisation has been linked to industrialisation and the growth of manufacturing employment and the shrinking importance of agriculture in a process known as the structural transformation.
A large body of literature within the field of economics depicts historical processes of urbanisation as the spatial outcome of a broader process of structural transformation which entails a reduction in the share of agriculture in employment as well as total value added as the economy expands (Chenery and Syrquin, 1975; Timmer, 2009). Globally this process has been under way for some time, with agriculture at present employing a third of the global labour force and producing a share of value added of 2-3 per cent (Satterthwaite et al., 2010). Standard structural transformation models focus on the interplay between rural and urban labour markets as drivers of urbanisation as increasing agricultural productivity expands food availability and gradually allows rural people to enter non-farm employment. Demand for services and trade links intensify with the upgrading of agriculture, resulting in stronger rural linkages to small towns, while urban demand for agricultural products increases as a result of urbanisation (Bairoch, 1988; Haggblade et al., 2007). These models revolve around a presumption of rising agricultural productivity, growing urban demand and increasing industrialisation.
In the case of Africa, however, urbanisation in the short run at least is predicted to be induced primarily by a set of negative rural pressures related to population growth, climate change, a declining natural resource base and generational conflicts over land, rather than the attraction of urban areas per se. Paradoxically, the continuing importance of agriculture is by necessity an inherent characteristic of African urbanisation, however, as livelihood opportunities outside agriculture are limited (Andersson Djurfeldt, 2015; Vandercasteelen et al., 2018). The distribution of the urban population in terms of size classes of cities and towns in many ways reflects this: the majority of the urban population lives in cities below one million inhabitants and 26 per cent in small towns (GOLD IV, 2017).
Urban transition and lower-level urban growth
For countries with large rural populations, small towns take on a special role as the outlets for localised growth dynamics within agriculture, both as outlets for agricultural commodities, and as centres of consumption and non-farm employment (Tiffen, 2003; von Braun, 2007). As such, small towns may be more strongly embedded spatially in their rural surroundings than in urban hierarchies, as posited, for instance, by theoretical perspectives inspired by the concept of subaltern urbanism (Roy, 2011).
The linkages between rural areas and lower-level urbanisation are recognised also in the integrated spatial-development perspectives that emerged in the 1980s, which resurrected earlier notions of trickle-down, spread effects and agglomeration economies found in the growth-pole perspectives of Perroux (1955). Rondinelli’s (1983; 1988) work on urban functions in rural development (UFRD) has perhaps been the most influential in this regard, but, like its precursors, it departs from the presumption that the rural can be developed from the urban. The concepts of territorial development and balanced growth acknowledge the interdependencies between urban and rural areas, albeit generally at a larger geographical scale (Tacoli, 2008; Barca et al., 2012).
More recent literature nuances the understanding of lower-level urbanisation in a number of respects. On the one hand, the distinction between push- and pull-based scenarios acknowledges that not all urbanisation is driven by dynamic change, but can also be related to stagnation and decline. As suggested by Haggblade et al. (2007), where access to agricultural assets is relatively equal and initial production potential is high, rising agricultural labour productivity is more likely to emerge, enabling family members to be pulled into nonfarm activities - pursuits which over time tend to be concentrated in urban centres. Under less conducive regional conditions, agricultural productivity falls as a result of diminishing natural resources, pushing households into labour-intensive, low-return non-farm activities. Urban areas under such conditions function as outlets for sale of products from these activities rather than centres of economic dynamism, with migration from rural areas being driven by a survivalist movement out of agriculture rather than a dynamic process of agrarian change.
Recent empirical research stresses the importance of rural context to the growth and development of small towns. Studies of secondary towns in Uganda and Tanzania suggest that both trade in agricultural goods and access to global trading networks focused on consumer items provide economic opportunities for particular towns (Mainet and Racaud, 2015). Lazaro et al. (2019) document how the historical evolution of particular value chains and diversification dynamics within agriculture affect the growth of rural settlements in Tanzania. While there is increasing recognition, on the one hand, of the importance of rural hinterland characteristics to lower-level urbanisation and, on the other, of the importance of contextualising such urbanisation in relation to national and global policies, there is limited understanding of how particular small towns are connected to broader urban systems. A further blind spot relates to the ways in which this interaction encourages or obstructs local-level processes of urban growth, both in the towns themselves and in their rural surroundings.
Here urban-systems theory (developed mainly within the context of the global North) provides interesting possibilities for theoretical hybridisation, especially in relation to recent empirical evidence from the global South that points to the importance of the composition of urbanisation to the poverty-reducing effects of the urban transition (Christiaensen et al., 2013).
Like growth-pole perspectives, urban-systems theory draws its heritage from Christaller’s (1933) central-place theory and the theoretical construct of a hierarchy of places based on population size and function. Later modifications of central-place theory have sought to combine this focus on hierarchy with general-systems theory to develop the concept of an urban system (that is, a group of urban centres connected through the movement of commodities and people), but have retained a monocentric focus on a paramount city (see e.g. Pred, 1977). In the case of the global North, as noted by Burger and Meijers (2012), spatial dynamics have changed considerably in the postindustrial era and more recent versions of urban-systems theory have come to place increasing emphasis on the concept of polycentricity. The latter considers both the morphological aspects of urban systems in which populations have become increasingly dispersed over urban centres of various sizes, but also functional aspects of connectivity and density of networks between urban centres (Green, 2007). Although urban-systems theory has been developed in the context of the global North, several of the concepts resonate with planning models and concepts applied to the global South, for instance the notion of balanced growth, and more recently that of territorial development (Tacoli, 2008; Barca et al., 2012).
Christiaensen and Todo (2014) and Christiaensen et al. (2013), using cross-country data for developing countries for the 1980-2004 period, show the comparative poverty-reducing effects of secondary town migration, but their analysis distinguishes only between metropolitan areas (above 1 million inhabitants) and secondary cities (the rest). Here urban-systems theory can inspire several new paths of inquiry. Delineating the morphological polycentricity that exists beneath the 50,000 population threshold is useful for understanding distributional patterns of urbanisation with respect to different size classes of small towns. Furthermore, analysing functional polycentricity and the connections as well as density of networks between different classes of small towns, larger urban centres and rural surroundings can explain the distribution of economic growth, poverty reduction and social and economic inclusion among different types of urban centres.
Theoretical perspectives on urban systems as well as empirical studies of secondary urbanisation focus on the broader spatial outcomes of urbanisation, for instance the emergence of regional disparities in economic growth. Neither of these bodies of scholarship considers aspects of inequality within urban centres, however. In this respect, therefore, there is limited knowledge of how the distribution of income, assets and opportunities among inhabitants of small towns relates to the position of particular urban areas within the urban system as a whole and their rural surroundings. Analysing the inclusivity of such processes from the perspective of gender and generation is a further challenge.
Decentralisation policies and local governance
Theoretical perspectives on lower-level urbanisation need to be contextualised in relation to the empirics of policies that may influence small towns as well as the livelihoods of their inhabitants. Decentralisation reforms have been promoted across the global South since the late 1990s, ostensibly to advance good governance as well as efficiency. Especially in geographically large and ethnically diverse countries the devolution of power to lower administrative tiers is tied to stronger stakeholder involvement, development policies better suited to local conditions and bureaucratic expediency. In this sense, decentralisation in itself has been presumed to be a generator of economic growth and development (World Bank, 2000).
The success of particular decentralisation programmes has varied strongly, however with Indonesia, the Philippines and to a lesser extent India put forth as positive examples, while the African experience with decentralisation has been stymied by problems of lacking resources and implementation capacity. The practical effects of decentralisation on particular cities and towns have, moreover, been largely related to the relative dependence on intergovernmental transfers compared with local revenue sources, but studies of such effects on municipal budgets are piecemeal and restricted to higher-level urban centres (GOLD IV, 2017). The broader economic consequences of decentralisation policies for small towns constitute an empirical gap, however. For instance, the freedom to locally set the agenda in terms of provision of urban services, manage local housing markets or govern livelihood opportunities through local bye-laws could provide explanations for both population growth and economic dynamism of particular urban centres (Andersson, 2002).
The case for a multi-scalar approach
A multi-scalar approach is necessary to analyse how lower-level urban centres are connected to wider urban systems as well as their rural hinterlands. Here urbansystems theory can be fruitfully combined with perspectives from livelihoods research. In broad terms, livelihoods research seeks to understand how households and individuals make their living through combining various types of capital. Drawing on the Sustainable Livelihoods Framework originally developed for rural areas by Chambers and Conway (1992) and modified by Scoones (1998) and Bebbington (1999), the livelihoods framework has also been adapted for urban livelihoods research, for instance, by Rakodi (1999). At the overarching level, the urban transition entails a movement away from a reliance on natural capital (land and water), towards livelihoods based increasingly on other types of capital.
A recent development of the livelihoods literature has focused on how capitals are accessed and combined over space, in what are termed multi-local or translocal livelihoods. Such translocality involves both rural-to-urban connections as well as intra-urban and intra-rural linkages and includes engagement in rural agriculture by urban households, transfers of food from rural to urban areas via family relations and the urban fostering of rural children. Gradually globalising labour markets, moreover, mean that translocality is also increasingly expressed through international remittances (De Haas, 2010). This literature connects to the broader literature on the multi-layered household systems found in the global South as providers of welfare and social insurance functions, generally delivered by the state under the welfare regimes of countries in the global North (Deshingkar et al., 2008; Andersson Djurfeldt, 2012; 2014).
The importance of access to translocal sources of livelihoods with respect to cash remittances and food security and the redistributive functions of such transfers have been shown by several researchers in an African context (Andersson, 2011a; 2011b; Crush and Caesar, 2018). This is in keeping with the notion that the broader family network serves to provide welfare functions for its weaker members. With respect to lower-level urban centres, the involvement in multi-local livelihoods can provide resilience to urban decline and stagnation through rural agriculture or the receipt of transfers from other urban areas - an explanation that may fly below the radar of official statistics. As such the density and nature of family networks may explain differences in social and economic inclusion between particular small towns. Nonetheless, as earlier studies suggest, kinship relations are based on norms and power structures that restrict opportunities and cement a division of labour based on gender, generation and social status also across space (Peters, 2019). The weaker tenure rights of women suggest that the drivers of migration as well as the opportunities for engaging in translocal livelihoods are also vastly different depending on gender and generation (Andersson Djurfeldt, 2019; Andersson Djurfeldt et al., 2019).
A nested approach aggregates the livelihoods of households and individuals into patterns of economic change at the level of small towns, but also considers translocal livelihoods as explanations for these patterns, including their inclusionary as well as exclusionary possibilities. The small town itself in turn is placed within the context of the urban system. This enables an exploration of how the characteristics of the broader urban system and surrounding rural areas relate to the intra-local distribution of assets and social and economic opportunities at the level of the individual town. Finally, this approach provides the possibility for analysing how variations in economic opportunities in lower-level urban centres can be explained by the ways in which they are embedded in national and regional settlement systems.
At the level of the urban system, the degree of polycentricity is theorised as conducive to broad-based patterns of urban growth. The translation of such growth into local-level inclusivity varies depending on the nature of economic activity (trade versus agriculture), the density of social networks and kin-based spatial linkages (strength and adaptability of kin relations), linkages to rural areas (land-tenure systems and engagement in rural agriculture), active policy measures (agricultural policies encouraging smallholder agriculture, cash transfer programmes and decentralisation policies) and permissive urban stances to informal economic and social activities.
Theoretically, therefore, this approach places not only the locations of urban transition in the context of both urban systems and rural surroundings, but also the livelihoods of the inhabitants of lower-level centres in relation to broader processes of demographic, economic and political change.
Academically, this nested approach therefore advances the research frontier through moving beyond perspectives that focus on urbanisation as the spatial outcome of a macro-level process of social and economic change. Rather it suggests that patterns of lower-level urbanisation are shaped not only by these processes but also by social and spatial relations embedded in these places through the lived experiences of their inhabitants. Lower-level urbanisation in this sense is built from the bottom up as much as conditioned by broader social and economic processes. In turn, this necessitates new empirical and methodological approaches that can capture the inclusivity of spatial patterns, economic processes and social relationships. Given such advances, a multi-scalar approach to lower-level urbanisation holds the promise for addressing the calls for social and economic inclusion found in policy agendas focused on sustainable urbanisation, for instance Agenda 2030 and the New Urban Agenda.