VoxEU Column Economic history Migration

Migration before railways: Evidence from Parisian prostitutes and revolutionaries

Little is known about migration to cities in the era before railways. The column uses data on the origins of women arrested for prostitution in Paris in the 1760s, women registered as prostitutes in the 1830s and 1850s, men holding identity cards during the French Revolution, as well as everyone buried in 1833 to examine patterns of migration. Migration was highest from areas with high living standards, and the impact of distance fell as transport improved. Distance was a stronger deterrent to females than to males, consistent with more limited employment opportunities for women.

With deaths rates that exceeded their birth rates, cities historically relied on migrants to fuel their rapid growth. In 1790s Paris, for instance, only 30% of the population had been born there. But because we lack reliable censuses before the middle of the 19th century, it is difficult to understand what drew migrants into these cities. However, in a recent paper we are able to examine the dynamics of working class migration into Paris between the mid-18th and mid-19th centuries (Kelly and Ó Gráda 2018). To do this we use information on two groups: female prostitutes and male holders of identity cards during the French Revolution. We supplement this with records of everyone buried in Paris in 1833. We find that migration at that time is strongly explained by a gravity model in which distance deterred women more than men, and that the impact of distance diminished as transportation networks improved. 

Parisian prostitutes 

Prostitution might not appear to be a promising avenue to understanding migration. Although prostitution has always been a survival strategy for women in dire economic circumstances, it has been morally taboo and usually illegal. This means that we usually do not know much about the women who worked as prostitutes. Paris is, however, a notable exception. 

Eugène Atget: Prostituées de la rue Asselin, 1925.

Before the Revolution, prostitution was formally illegal and records of the women arrested in Paris during the 1760s – their ages, their previous occupations, and where they had come from – formed the basis of a classic study by Benabou (1987). After 1789, the state decided that prostitution, being inevitable, should be regulated rather than prohibited. Women working as prostitutes were required to register with the police and undergo regular health checks (Harsin 1985). Alexandre Parent-Duchâtelet, a pioneer of public health and social statistics, collated these registration data. He supplemented this with other information based on frequent visits to brothels (“always accompanied by an inspector”), and extensive interviews with prostitutes in De la prostitution dans la ville de Paris (Parent-Duchâtelet 1836). This is still the most systematic study of female sex workers ever undertaken. It contains extensive accounts of their lives, and describing the organisation and clientele of the trade in brothels, lodging houses, and on the streets. 

Throughout the book there are two contradictory attitudes to prostitutes. On one hand, Parent-Duchâtelet argues that the innate degeneracy of these women posed a fundamental threat to the social order, a threat that could only be contained through vigorous regulation. His empirical findings were in tension with this opinion of prostitutes as social pathogens. He found that almost all Parisian prostitutes were young working-class women, typically illiterate, who had been driven into prostitution by their economic circumstances. 

For instance, in the middle of a lengthy discussion of how women became prostitutes through an aversion to work and a love of luxury, the author presented a table of the reasons that the women themselves gave. In equal numbers, they were destitute, had been orphaned or expelled from home, or had been 'kept women' who were discarded by their lovers (or fled abusive ones) and so were in effect moving from unregistered to registered prostitution. Parent-Duchâtelet noted that low wages and precarious employment left many vulnerable to the choice between prostitution and starvation – and then returned to his theme of how these women had progressed through disorderly lives to debauchery and finally to prostitution. 

He reported that most of those who registered were aged between 16 and 25, with a modal age of 20. Among these registered prostitutes, 3% were children aged between 10 and 15. This was technically illegal, but children judged to be irremediably debauched were sometimes registered. 

Few women remained long in the occupation – only 43% were still registered after four years, and 36% after five. Most were uneducated, and two-thirds had previously worked in textiles or apparel, with most of the rest in some form of retail. Most women from the provinces were the daughters of artisans, labourers or farmers. 

Unsurprisingly, these characteristics mirror what Henderson (1999) found in court records for 18th century London. Again, London prostitutes were almost exclusively women aged around 20 from the lowest levels of society, and rarely worked as prostitutes for more than five years. van de Pol (2012) found the same pattern in 17th century Amsterdam. 

Trafficking or coercion was not recorded as a reason for entering prostitution in Parent-Duchâtelet's research. This is possibly because widespread destitution provided an adequate supply of women into the trade. Parent-Duchâtelet’s book, however, was concerned only with lower-class prostitution. He did not consider the demimonde of dames entretenues and high-class brothels catering to the wealthy. Kushner (2016) showed, using the files of the Département des femmes galantes – a branch of the secret police specifically devoted to the surveillance of courtesans as a means of gathering information on their upper-class clients – that coercion was common in the late 18th century. Girls of 14 and 15 years old were routinely sold by their parents to madams or individuals, with the full knowledge and acquiescence of the police. 

Figure 1 Origin of female migrants to Paris, 1760s-1854

Source: Kelly and Ó Gráda (2018).

Female migration 

We want to understand the factors that drove migration, especially the impact of distance and living standards. Formally, we estimate a gravity model of migration, something that originated with Ravenstein (1885) and was formalised by Zipf (1946). The share of a region’s population that migrates to another region should depend on distance and relative living standards. Living standards can affect migration in two ways. One one hand, poverty increases the rewards to migration. On the other, people living in more affluent areas with higher literacy are better able to afford the fare and are more likely to find work in the city when they arrive. 

Besides information on the origin of prostitutes in the 1760s, 1834 and 1854, we have data on the origin of all women buried in Paris in 1833, who would mostly have arrived sometime around 1800. Figure 1 maps the places in which these women originated. The pattern of migration changed markedly between the 1760s and 1800, if we assume migration in 1800 is represented by the burials in the second panel. Migrant origins were concentrated to the east of Paris in 1760, but later were spread more uniformly in all directions. Between 1800 and 1834 the pattern of migration was effectively unchanged. By 1854 the impact of distance had lessened, as Figure 2 shows. 

These changes in migration patterns match the evolution of the French transport network. Between 1765 and 1780, an ambitious scheme of road improvements, using improved construction techniques and coerced labour, more than halved the journey time between Paris and most provincial cities (Clout 1977). In the 1840s, railway networks began to appear and, by 1850, railways connected all major cities to Paris. The network more than doubled in length in the following decade. As a result, the elasticity of migration with respect to distance fell from -1.5 for burials in 1833 and prostitutes in 1835, to -1.2 in 1854. At the same time, higher living standards, measured by the number of army recruits who were literate, markedly increased migration. 

Figure 2 The declining impact of distance on female migration, 1834-1854 for registered prostitutes in Paris by birthplace 1834 and 1854 

Source: Kelly and Ó Gráda (2018).

Male migration

For men, besides 1833 burials, our main sources of information are the identity cards issued to all men by the Revolutionary government in 1792–1793, as compiled by Blum and Houdaille (1986), and records of recruits into Napoleon’s army from 1801 to 1814 (Dupâquier and Goy 1989). The fall-off of migration with distance was less than for women. The gravity coefficient was only -1 (Figure 3), and the impact of living standards measured by literacy was considerably less. This is consistent with the idea that the employment opportunities for men, in construction and manufacturing, were much better than those for women. Among unmarried people in their twenties living in Paris in 1851, men outnumbered women two-to-one (Conner 2017), an imbalance that Corbin (1990) has argued was a large factor in driving the market for prostitution. 

Figure 3 Distance as a deterrent to female and male migration, burials in Paris, 1833

As railway networks expanded, the tyranny of distance shrank. The 1891 census implies a gravity coefficient of only -0.6, and no impact of local living standards, but its table on birthplaces did not distinguish women from men. For Marseilles, we find that registered prostitutes in the 1880s had the same gravity coefficient as the census population. Brides had become as likely as grooms to have come from places outside the city’s hinterland (Sewell 1985). This suggests that the patterns of female and male migration had become broadly similar by the late 19th century. 


Benabou, E-M (1987), La prostitution et la police des moeurs au XVIIIe siècle, Librairie Académique Perrin. 

Blum, A and J Houdaille (1986), “12,000 Parisiens en 1793. Sondage dans les cartes de civisme”, Population: 259–302. 

Clout, H D (1977), Industrial Development in the Eighteenth and Nineteenth Centuriesin Themes in the Historical Geography of France, Academic Press. 

Conner, S P (2017), "The paradoxes and contradictions of prostitution in Paris", in M Garcia, L H van Voss, and E van Nederveen Meerkerk (eds), Selling Sex in the City: A Global History of Prostitution, 1600s-2000s, Brill. 

Corbin, A (1990), Women for Hire: Prostitution and Sexuality in France after 1850, Harvard University Press. 

Dupâquier, J, J Goy (1989), "Révolution et population", in J Dupâquier (ed.), Histoire de la population française, vol. 4,  Presses Universitaires de France. 

Harsin J (1985), Policing Prostitution in Nineteenth-Century Paris, Princeton University Press. 

Henderson, T (1999), Disorderly Women in Eighteenth Century London, Longman. 

Kelly, M and C Ó Gráda (2018), "Gravity and Migration before Railways: Evidence from Parisian Prostitutes and Revolutionaries", CEPR Discussion Paper 13046.

Kushner, N (2016), Erotic Exchanges: The World of Elite Prostitution in Eighteenth-Century Paris, Cornell University Press. 

Parent-Duchâtelet A J B (1836), De la prostitution dans la ville de Paris, considérée sous le rapport de l'hygiène publique, de la morale et de l'administration, Baillière et fils.

Ravenstein, E G (1885), “The Laws of Migration”, Journal of the Statistical Society of London: 167–235. 

Sewell, W H (1985), Structure and Mobility: The Men and Women of Marseille, 1820–1870, Cambridge University Press. 

van de Pol, L (2011), The Burgher and the Whore: Prostitution in Early Modern Amsterdam, Oxford University Press. 

Zipf, G K (1946), “The (P1P2)/(D) Hypothesis: On the Intercity Movement of Persons”, American Sociological Review: 677–686. 

2,310 Reads