Urban life has been studied within a variety of disciplines – anthropology, sociology, politics, architecture, geography, and economics. Scholars have used urban analysis for diverse purposes, from predicting troughs in the business cycle by observing skyscrapers (Thornton, 2005) to writing “personal” biographies of individual cities – most recently Jerusalem (Sebag Montefiore, 2011). Within economic history, the study of cities focuses on the relation between urbanisation and development, and more specifically on the study of factors such as employment, wages, disposable income, access to capital and public goods, as well as more general economies or diseconomies of urban density and agglomeration of industry.
Over the last two centuries, the world has undergone an unprecedented and rapid process of urbanisation: Less than 10% of the population of England, and probably less than 3% of the world’s population, lived in urban settlements in 1800 (Hoyt, 1963: 179) and today more than 50% of humanity lives in cities (UN DESA 2009). While the move towards urbanisation has been universal, it was (and still is) characterised by discrete features across geographies and historical periods. The variance in key drivers, pace, composition, and, indeed, outcomes of urbanisation seems to divide, crudely, into two eras, each with its own protagonists and characteristics. As Clark points out, “urban development was largely confined to developed countries” before the middle of the 20c, but has since “spread to developing countries” (1998: 85). In this short essay, we will provide an overview of the theory concerning the drivers and patterns of urban development within the context of economic history; proceed with a thematic examination of several factors – employment prospects, government policy, technology, mortality and fertility – affecting the development of urban centres; and conclude with a summary of the reasons for urban growth in the second half of the 20c in comparison to the previous era.
The story of urbanisation is intertwined with the story of global industrialisation, and one is often described as the driving force behind the other: Development economists such as Simon Kuznets and Arthur Lewis saw urbanisation as a “positive force in economic development” and “an integral part of economic growth and distributional change (including poverty reduction) in poor countries” (Ravallion et. al. 2007:667); Acemoglu, Johnson, and Robinson relied on population density to assert that urbanisation can serve as a reliable proxy for income per capita in their historical study of institutions, geography, and income distribution (2002: 1242-43); and Maddison, who compiled time series estimating GDP per capita across the globe over the last three centuries, pointed out that people in countries with higher urbanisation rates tend to trade more with each other, with a higher proportion of that trade being in services (Maddison 1983: 37 and footnote no. 12.). Other observers noted that while the world’s poor were indeed once “huddled largely in rural areas”, they have since “gravitated to the cities” (1997:58). Yet, following a survey of living conditions in 90 countries, Ravallion et al still estimate that “about three-quarters of the developing world’s poor… live in rural areas” (2007: 693). This assessment, however, relies on a specific set of definitions of both poverty and urbanisation. Indeed, the study of urbanisation and development is overshadowed by several methodological challenges including the lack of reliable comparable data, a lack of a common standard definition for an “urban” area and of poverty1, and the lag in availability of data (Hohenberg & Hollen Lees, 1995: 218)*.
While the jury is still out on the exact relation, if any, between urbanisation and industrialisation, the scholarly interest in this subject over the past 60 years seems to be among the causes – perhaps a symptom – of one of the unique features of urban growth in late-developing countries: A growing attempt to plan and control urban development based on one theory or the other, often with severe unintended consequences. We will go over some examples of this phenomena below, but first let us summarise the less controversial aspects of the relevant theories concerning urbanisation. In Principles of Economics (1920), Alfred Marshal provided “the first careful economic analysis of agglomeration economies, arguing that cities enhance productivity by allowing for labor market pooling, input sharing, and technological spillovers” (Rosenthal & Strange, 2003:377). And indeed, a variety of empirical studies found “positive correlation between agglomeration and aggregate growth” (Martin & Ottaviano 2001: 967), and elaborated on the positive human capital externalities associated with urban industry (Ciccone & Perri 2005). Still, it would be wrong to conclude that growth in urbanisation, for better or worse, always goes hand in hand with economic development or industrialisation. As Hohenberg and Hollen Lees (1995) point out, urbanisation pattern in now-developed countries only started to show a clear link with industrialisation around the time of the 2nd Industrial Revolution in the late 19c, following two phases of urban growth driven by (1) proto-industrialisation surrounding existing urban centres and (2) proximity to natural resources. Looking beyond the relation between urbanisation and development, city life was historically associated with a variety of negative factors, broadly known as the “urban penalty”, which we will touch upon in our discussion of population growth below. Overall, the existing theory provides a useful framework for ex post analysis of development but not a predictive, or even prescriptive, model.
We now turn to examine the growth of urban centres in the second half of the 20c. In 1970, only 25% of the population of the world’s less developed regions lived in cities, compared to close to 40% by the end of the millennium; for the world’s more developed regions, the figures were 67.5% and close to 90% respectively (Pacione, 2009: 71). This increase in the percentage of urban living corresponded with a virtual doubling of overall global population. Cities are also becoming larger: As Pacione points out, “the average population of the world’s largest cities was over 5 million inhabitants in 1990, compared with [only] 2.1 million in 1950, and less than 200,000 in 1800”. The number of cities with 8 million or more inhabitants is also increasing rapidly, particularly in late developing countries (p.74). The rise of the late-developing “megalopolies” is apparent when examining the list of the world’s largest urban agglomerations: while in 1950 the majority of the top 15 cities on the list were in relatively developed countries, by 2000 the only developed survivors on the top 15 were Tokyo, Osaka, New York, and Los Angeles, as well as Seoul, the capital of South Korea, which is now considered a capital of a developed country.
As we alluded to above, urban migration was historically associated with superior employment possibilities. And, as Todaro and Smith points out, rural-urban was still “primarily an economic phenomenon” in the second half of the 20c (2000: 345). However, while urban industry generally affords markedly higher wages, it is also plagued with a relatively high unemployment rate and thus provides a mixed basket of incentives for migration. Nonetheless, it is still attractive compared to rural employment and, as the Todaro model suggests, “proceeds in response to urban-rural differences in expected income [sic.] rather than actual earnings”, meaning that the “individual must balance the probabilities and risks of being unemployed or being underemployed for a considerable period of time against the positive urban-rural real income differential” (Todaro & Smith, 2000: 345-6). And so, the model assumes, rural-urban migration might “continue even if the urban unemployment rate were 30% to 40%” (p.347). The prevalence of high urban unemployment in the second half of the 20c – often described as a market failure – is partially explained by the theory of Efficiency Wages, an attempt to model the way in which firms respond to changes in demand by cutting their workforce and not by reducing salaries (Weiss, 2008), and governments have tried to alleviate this tendency in a variety of means including wage subsidies, production subsidies and tariffs (Chin, 1998: 295).
The relatively high unemployment rate in 20c urban areas is associated with a large informal sector, another unique feature of late developing countries. As La Porta and Shleifer points out, “in developing countries, informal firms account for up to about half of all economic activity”, which carries with it a variety of negative implications for health and safety, insurance, tax evasion, and lower overall productivity (2008: 275). In consequence, “the size of the informal economy is strongly negatively correlated with income per capita” (p. 284), and young labor from rural and semi-rural areas often forms the low-earning strata of the urban labor market. This is the case in India (Shaw & Pandit, 2001:180), China (Wong et al, 2006), as well as in other developing countries. While migration to now-developed countries in the 19th and early 20c was largely a family affair, in many LDCs it was often driven by individual migrants, with gender being a significant factor in driving and shaping migration patterns: In the Philippines, to cite one example, the “notion that migrants tend to enter the urban occupational structure through the informal sector seems to apply mainly to female migrants” (Koo & Smith, 1983: 219) in line with broader cultural and historical drivers on inequality in Asia (Bauaer et al, 1992); in Latin America, to cite another example, “practices and beliefs in the home and community…. construct gender ideologies which undervalue the work of women. Enterprises then take advantage of the cheapness of women’s labour and so reinforce the ideology of undervaluation” (Watson in Gilbert, 1994: 619); and so, the general inequities and inefficiencies of the informal sector are exacerbated. In addition, advances in banking and communication technologies make it easier for families to live on the edge of the urban world, with one of the spouses working in the city and supporting the other members through remittances. This has become one of the common features of reform-era China (Li et al 2001: 205; Qian, 2003; ).
Beyond individual perceptions and biases, government policy played a significant role in shaping late urbanisation, both directly and through path dependencies associated with colonial rule. As Preston (1979) points out, colonial heritage was a key driver of “giantism”, a bias towards larger cities, in Latin America, and the extreme coastal concentration of some of the region’s large cities can be explained by “trade relationships and natural resource exploitation by colonial powers” (Hardoy in Preston 1979:200). Such colonial “mementos” are not unique to latin America; the overdevelopment of Jakarta, one of the Asia’s largest cities, was previously promoted by the Dutch who concentrated their administrative bureaucracy in the city, a tendency that was later “exacerbated under local rule” (Hugo in Preston 1979:200). As we noted above, studies and perceptions linking urbanisation with development also contributed their share, driving and justifying “exaggerated bias of government expenditures on infrastructure and services in favour of urban areas” (Clark 1998:93), perhaps in the hope that urbanisation will bring about development. On the other hand, China, for example, limited internal migration by enforcing a dual-system of urban and rural social services. More broadly, the general increase in government involvement in economic and social affairs, and the authoritarian nature of many of the world’s LDCs meant that growth in the second half of the 20c was less organic and dispersed than in the previous era and more prone to unintended consequences. This includes a marked bias towards larger of ‘first-place’ cities who received “a disproportionately large share of public investment and incentives for private investment” (Todaro & Smith 2000:332). Apart from government attempts to shape and control migration, urban congestion suffered “collateral damage” from other policies that affect the economy as whole, such as attempts to promote more value-added production, and promotion of specific industries (Chen, 2006).
Some have suggested that the factors driving the growth of urban centres may soon be undermined by a powerful counterforce — technology. In The Death of Distance (1997), Frances Cairncross describes a future in which cities will become entertainment and civic activity centres and governments in “poor countries will stem the flight from the countryside… by using low-cost communications to provide rural dwellers with better medical services, jobs, education, and entertainment” (p. XV). Others, in contrast, pointed out that while some of the advantages of cities as centres of industry are no longer as pronounced, “agglomeration economies will continue to be large… information spillovers will continue to be important and telecommunications may end up helping, rather than hurting, cities” (Glaeser, 1998:157). The question is far from answered, but circumstantial evidence suggests that even the most “virtual” industries such as online services and telecommunications tend to concentrate in, and often drive the growth of, urban centres, wether it is in Silicon Valley, Bangalore, Hangzhou, or Tel Aviv.
So far, our analysis concentrated on factors that affected migration to urban areas; before we conclude, we must turn our attention to a more internal driver of growth. As we noted in the introduction, cities were historically associated with several negative externalities that adversely affected human health. These include(d) air and water pollution and a variety of diseases that thrived on population density. However, while in the 19c urban life was markedly shorter and more disease-prone relative to rural life, the turn of the century saw a significant decline in urban death rates (Schofield et al, 1999; Hubbard 2000), and, more recently, “the concentration of people and resources in cities, has been a dominant [positive] influence on health” (Freudenberg et al 2005: 1). In China, for example, mortality declined significantly in the second half of the 20c, in line with dramatic increases in urbanisation (Banister & Hill, 2004), and in the UK, to cite an example from a developed economy, a recent study found that urban mortality rates in urban and rural areas are relatively similar, albeit with significant differences in causes and following certain adjustments (Gartner et al, 2008). And so, while urbanisation is not necessarily a driver of declining mortality rates, improvements in sanitation and medical services mean that organic growth played a more significant role in recent decades than 100 years ago.
In conclusion, the growth in the world’s major urban centres in the twentieth century was driven by unique factors that differed considerably from those driving urbanisation in the 19th and early 20c. These include differences in employment prospects and perceived benefits, driven in part by distortions due to more competitive industry; path-dependencies from the colonial era; a more aggressive role played by government, characterised by direct policies to promote urban development and specific urban centres as well as by broader policies to promote urban-related industries; migration patterns, revolving around individuals and specific gender roles and less about families; and a somewhat more organic growth pattern following advances in sanitation and medical services. Since WWII, technology has also played a limited, if increasingly important, role in shaping urbanisation, but its influence does not yet lend itself to theoretical generalisations. The drivers of urbanisation in late developed countries have, in turn, affected overall development as well as brought to the fore each country’s failures and successes.