Why did urbanization increase in the last decades of the nineteenth century?

The story of nineteenth century development in the United States is one of dynamic tension between extensive growth as the country was settled by more people, bringing more land and resources into production and the intensive growth from enhancing the productivity of specific locations. Key elements in that intensive growth are the movement of population out of agriculture and into other activities. These activities included not only trade (with its closely associated exchange and movement of goods and people) which concentrated in marketplaces but also and increasingly over the century, manufacturing. This paper seeks to clarify the interaction of these forces with particular attention to the growth of population centers and their increasing colocation with manufacturing. At its inception, America’s first industrial revolution was predominantly rural. It would become increasingly urban over the course of the nineteenth century, not only as economic activity agglomerated around early nuclei but also because changes in technology diminished the importance and value of certain initial conditions and freed manufacturing to relocate to more centralized and concentrated locations. We document the increasingly urban character of manufacturing over the century, and we sketch an economic framework for its interpretation.

In 1800 83 percent of the American labor force was engaged in agricultural production, mostly for household consumption with any surplus of most crops directed towards local markets (Carter et al., 2006, Ba814 and Ba817). There were some American products where the principal market lay overseas, like tobacco and cotton, but these were the rare exception. Most basic manufactured goods were also made in the home although some might be purchased from local artisans. Household production was driven both by a lack of the necessary cash to buy in the market and the distance to readily accessible marketplaces.

Population densities in the United States were very low, especially when compared with those in western Europe or the Far East. There were few towns and cities of any size. In 1800, for example, the US population of 5.3 million was spread out across over 860,000 square miles of territory, giving the country an average population density of 6.1 per square mile—barely over the 1890 census threshold of “settled” (density ​> ​6.0).1 Moreover, the Louisiana Purchase (1804) approximately doubled the land area of the US, adding so much sparsely populated territory that despite the US population growing by 36 percent between 1800 and 1810, density fell to just 4.3 persons per square mile (far short of one family) and did not surpass the 1800 level until sometime in the 1820s (Carter et al., 2006). Just 33 communities nationwide in 1800 would have met the conventional Census definition of “urban” – population of 2500 or more – comprising 6.1 percent of the population.2 Only two cities (Philadelphia and New York) where 1.9 percent of the US population lived, were of any real size and would have met the population cutoff to qualify as one of largest 100 cities as of 1900 (that is, a population of 38,000 or more, see Gibson, 1998).

Moreover, the location of the few urban places in 1800 reflected the primitive nature of overland transportation, which was extremely limited and costly, even when it occurred on “improved” surfaces like turnpikes or graded and paved roads in or near the few population centers. Shipping goods by wagon over indifferent or bad roads – the vast majority -- cost as much as 70 cents per ton-mile in 1816 (Taylor, 1951), ruling out all but the least bulky, most valuable commodities. Personal travel overland was equally difficult, expensive, and time-consuming. While a horse might carry a rider 20–30 miles in a day, horses were costly, and most trips were made on foot. If a wagon were needed, a trip to and from market could easily take several days (Rothenberg, 1981). Much transportation, therefore, of goods or people in the early nineteenth century occurred along the Eastern seaboard or on navigable internal waterways. Fig. 1 maps the locations of urban places and geographic variation in population density in 1800.

The few relatively large agglomerations on the coast – Boston or New York, for example – were located where natural advantage provided harbors with ocean access. Further inland, water transportation had to stop where rivers draining to the Atlantic Ocean crossed the Eastern fall line, again creating points of agglomeration and concentrated settlement – so-called “portage” sites (Bleakley and Lin, 2012).3 However, as noted, the number of such places and the share of the population living in them were very small at the start of the nineteenth century – the vast majority of the population was rural, and sparsely settled at that.

Over the nineteenth century, the American economy grew dramatically in the aggregate and per capita, coincident with a pronounced shift of labor out of agriculture (Gallman and Rhode, 2019). By 1900, the share of the labor force in farming had decreased 43 percentage points during the preceding century. And, as the labor force shifted out of agriculture, the nation became more urban with almost 40 percent of the American population living in places of 2500 population or more and average population density increased. Moreover, fully a quarter of the population lived in the nation’s 100 largest cities (population of 38,000 or more). These cities are mapped in Fig. 2, along with the geographic variation in population density:

Many cities were deep in the interior, far from either coast and the eastern fall line or, in the case of the Midwest, the Great Lakes. Some had limited or no access to navigable waterways (for example, Indianapolis or Salt Lake City).

An important component of the shift of labor out of agriculture was into manufacturing. When the nineteenth century began, the Industrial Revolution was well underway in Britain but barely in its infancy in the United States (United States. Congress. House, 1791). By the early twentieth century, the United States was the world’s leading producer of manufactures, with labor productivity twice as high as that in Britain, the nation where industrialization had first taken hold (Broadberry and Irwin, 2004, Table 3; Broadberry and Irwin, 2006). Critical to American industrial ascendancy were radical changes in the nature of the manufacturing process and the rise of the factory. Home manufactures declined precipitously, and many local artisans were displaced by the factory system. Compared with artisan shops, factories employed more workers per establishment and used powered machinery to produce a large volume of output that was sold in increasingly distant markets. Early in the nineteenth century waterpower was the preferred source, but by the end, steam power predominated.

To make this transition work, inputs had to be purchased and product shipped all over the nation using a dramatically expanded and improved transportation network – the so-called “Transportation Revolution” (Taylor, 1951). The most important innovation was the development of a nationwide railway network. The diffusion of railways began in the 1840s and was essentially complete by the end of the century, dramatically increasing market access (Taylor, 1951; Fogel, 1964; Fishlow 1965; Atack, 2013; Donaldson and Hornbeck, 2016). Overall, the transportation revolution was associated with profound changes in the cost of transportation both within and between modes. Wagon haulage experienced the smallest change with per ton-mile rates perhaps halving to 15 cents where there were surfaced rather than dirt roads. Water rates, which were already low relative to overland rates, declined perhaps 30 percent on downstream passage; 66 percent transatlantic and 85 percent upstream to fractions of a cent per ton mile thanks in part to steam power. And, in railroading, productivity growth drove down shipping rates by 80 percent to the point where shipping by rail was cost competitive with rivers on a ton-mile basis but for much quicker and more direct delivery. The transportation revolution (especially the railroad) fueled the growth in the number and size of urban places. Coupled with the changes in manufacturing production, this incentivized establishments to urbanize. While it would be incorrect to say that such agglomeration could occur almost anywhere, sources of natural advantage in the early nineteenth century such as coastal access or portage, were much less important by 1900, as shown in Fig. 2.

Over the past several decades, American economic historians have assembled empirical evidence that allows this explanation of why manufacturing urbanized to be refined in ever greater detail. Some involves the digitization and processing of published data from the various nineteenth century census volumes, as well as archival evidence from the original manuscript census returns. It also involves digitization and GIS processing of historical maps of various kinds, part of the ongoing big data revolution in economic history. Similar efforts are underway for other countries experiencing historical industrialization and urbanization. We mention these briefly in the conclusion but focus mainly on research on the United States along with suggestions for improvement and further analysis moving forward.

Cited by (5)

• Can robots replace human beings? —Assessment on the developmental potential of construction robot

  2022, Journal of Building Engineering

  Show abstractNavigate Down

  Construction robots are a powerful means of addressing labor shortages and productivity inefficiencies. Studies discussed the application of construction robots on the level of economic and technical feasibility. However, the substitution potential of different construction robots has not been specifically explored in the context of the current situation of construction machine substitution. Task competence is the main reason for “machine substitution.” Thus, this research focuses on task competence attributes and builds an analytical framework for discussing human–task–machine in the construction field based on the basic framework of task-technology fit theory. Based on this analytical framework, this research sorts out the task processing logic of construction workers and robots. Moreover, this research establishes a measure of the substitution potential of “machine substitution” in the construction field. Taking the construction market of Mainland China as an example, this research analyzed the development potential of different construction robots, considering the current situation of “machine substitution” in the construction field. Results show that there is a shortage of construction robot types at present. For existing construction robots, differences exist in terms of how construction robots perform tasks competently and the degree to which they can replace construction workers. In general, the construction field is in a phase of human–machine collaboration and will remain so for a long time to come. Results can provide guidance for the application promotion and decision-making of construction robots.

• Can Public Housing Trigger Industrialization?

  2022, Journal of Housing Economics

  Show abstractNavigate Down

  The impact of public housing plans on local development is a neglected, although potentially important, issue. Here, we consider the impact of a public housing supply shock in a spatial equilibrium model and show that a larger local availability of houses can trigger industrialization by raising the number of residents. Also, the model suggests that this mechanism is stronger in places that exhibited, prior to the public housing plan, relatively higher population density. These implications are confirmed by the evidence we find from the INA-Casa plan, a program implemented by the Italian government in the aftermath of WWII.

• Introduction to the Special Issue: “Urban economics and history”

  2022, Regional Science and Urban Economics

• History and urban economics

  2022, Regional Science and Urban Economics

  Show abstractNavigate Down

  This article reviews recent literature using insights from history to answer central questions in urban economics. This area of research has seen rapid growth in the past decade, thanks to new technologies that have made available increasingly rich data stretching far back in time. The focus is to review innovative methods to exploit historical information and discuss applications of these data that provide new insights into (i) the long run growth of cities or regional economies and (ii) the spatial organization of economic activities within cities. The review also surveys the growing literature outside urban economics that uses the historical urbanization as a proxy for economic growth, discusses differences between how economic historians and urban economists think about the relationship between urbanization and growth, and considers how these views might be reconciled.

• Spatial Dynamics and Determinants of Population Urbanization in the Upper Reaches of the Yellow River

  2022, Land

Recommended articles (6)

• Research article

  Urban issues in Health Economics: Guest editors' introduction

  Regional Science and Urban Economics, Volume 93, 2022, Article 103749

• Research article

  Organization & density-related differences in within-firm wage disparities

  Regional Science and Urban Economics, Volume 95, 2022, Article 103787

  Show abstractNavigate Down

  This paper examines differences in within-firm wage disparities across French employment areas. Using administrative data, my analysis first documents that wage disparities are larger within firms operating in denser employment areas: a 100% increase in the employment density of an area is associated with a 0.015 log point increase in the 95/5 hourly wage ratio within firms. My analysis also examines the extent to which differences in the internal organization of labor into layers of hierarchy account for these disparities. I find that firms in denser employment areas organize into a greater number of layers, and that wage disparities are larger in firms with more layers. Finally, I conclude that firm organization is important: local differences in firm organization account for 31.3%–50.0% of the cross-sectional relationship between the employment density of areas and the log of the 95/5 hourly wage ratio within firms, and 12.2% of the same relationship estimated within firms. A separate analysis conducted with the standard deviation of log hourly wages leads to similar conclusions.

• Research article

  The relative importance of portable and non-portable agglomeration effects for the urban wage premium

  Regional Science and Urban Economics, Volume 95, 2022, Article 103786

  Show abstractNavigate Down

  Using administrative data for West Germany, we study the relative importance of portable and non-portable agglomeration effects for the urban wage premium. In doing so, we advance the established strategy of estimating wage-tenure profiles for urban-rural and rural-urban movers by adding worker, firm, and match fixed effects. This allows us to distinguish unambiguously between both types of agglomeration effects. Our results show that portable and non-portable agglomeration effects equally contribute to the urban wage premium. Moreover, portable agglomeration effects are not only observed in the biggest cities. Instead, the speed of human capital accumulation continuously increases with city size.

• Research article

  Location of housing and industry around city centre amenities

  Regional Science and Urban Economics, Volume 95, 2022, Article 103739

  Show abstractNavigate Down

  We investigate the endogenous emergence of residential and industrial areas when proximity to the centre benefits both households and firms. Households wish to locate near firms and near the centre due to costs of commuting and of visiting the centre that are proportional to distance. Firms wish to locate near households because this reduces the wage they need to pay to attract workers, and near the centre due to the iceberg cost of transporting their output to the centre. Depending on the relative magnitude of three cost parameters (commuting to work, visiting the centre, transporting output to the centre), different urban structures arise. If the cost of visiting the centre is large and the cost of commuting to work is small (relatively to the cost of transporting output), households locate in the centre and firms in the outskirts. The opposite occurs if both costs are very small. If the cost of visiting the centre is small and that of commuting to work is large, housing and industry become completely integrated. Complex urban structures emerge in intermediate cases.

• Research article

  Geographic and temporal variation in housing filtering rates

  Regional Science and Urban Economics, Volume 93, 2022, Article 103758

  Show abstractNavigate Down

  In the field of Housing Economics, filtering is the process by which properties, as they age and depreciate in quality, tend to be occupied by lower-income households. This is the primary mechanism by which competitive markets supply low-income housing. While filtering is an important long-term source of lower-income housing at the national level, this research shows that filtering rates for owner-occupied properties vary considerably both across and within metropolitan statistical areas. Notably, in some markets, properties “filter up” to higher-income households. This paper contributes to our understanding of filtering by demonstrating the geographic and temporal heterogeneity of filtering rates and examining links between filtering, supply elasticity, and gentrification. We also explore two alternative measures of filtering based on changes in relative income rather than real income.

• Research article

  On the economic geography of an aging society

  Regional Science and Urban Economics, Volume 95, 2022, Article 103798

  Show abstractNavigate Down

  This study examines the spatial distribution of young and old consumers and the impacts of the aging population on economic geography. We pay special attention to the generational difference in the incentive behind a location choice, and the interplay of agglomeration and dispersion forces. The results show that the elderly exhibit a strong tendency to agglomerate, and the increase in the aged population promotes the spatial agglomeration of economic activities. In addition, it is shown that urban costs weaken the tendency toward agglomeration. This study is one of the first attempts to combine a new economic geography model and a discrete-time overlapping generations model.