Railroads, specialization, and population growth: evidence from the first globalization


We explore how railroads affected population growth during the first globalization (1865–1920) in Chile. We look at areas with a strong comparative advantage in agriculture using novel data that document 60 years of railroad construction. Using instrumental variables, we present four main findings. First, railroads increased both urban and rural population growth. Second, the impact was stronger in areas with more potential for agricultural expansion. Third, railroads increased specialization in agriculture when combined with a high level of the real exchange rate. And fourth, railroads had little effect on human capital and fertility. These results suggest that the effects of transportation technologies depend on existing macroeconomic conditions.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8


  1. 1.

    Even today, projects like the construction of the “Belt and Road Initiative” implemented by the Chinese government incorporate the construction of railroads (Brakman et al. 2019). Fogel (1962) and Fogel (1964), Fishlow (1965), and Rostow (1967) were pioneers in studying railroads in economics. Similar technologies include the steamship (North 1958) and the refrigerator car (Kujovich 1970).

  2. 2.

    Fishlow (1965, p. 203) puts it simply: “whether railroads first set in motion the forces culminating in the economic development of the decade, or whether arising in response to profitable situations, they played a more massive role.”

  3. 3.

    Katz (2018) also applies the idea of “straight lines” in a panel data context for the USA using a different approach, which exploits the appearance of new big (connecting) cities.

  4. 4.

    Our identification strategy is suited for the center and south parts of the country and it is not valid for the north, the reason why we exclude it from the analysis in the paper.

  5. 5.

    We construct the potential for agricultural expansion using data on agricultural suitability (FAO-GAEZ), also used by others in the literature (e.g., Alesina et al. 2013; Galor and Ozak 2016).

  6. 6.

    We argue that the real exchange rate was exogenous to the process we study and driven by mining activities in the north (Badia-Miro and Ducoing 2015; Humud 1974; Mamalakis 1976; Meller 1998).

  7. 7.

    We implement a battery of exercises to confirm that the mechanism of relative prices is behind the estimated effects and not alternative channels related to other macroeconomic variables.

  8. 8.

    Other papers studying the causal effects of railroads in Latin America are Miller (1976), Coatsworth (1979), Ramírez (2001), Summerhill (2005), Herranz-Loncán (2014), Zegarra (2011), and Perez (2018).

  9. 9.

    The first railroads were constructed in the north to fulfill the demand of the mining industry. William Wheelwright, an American entrepreneur, led that process. The construction of these railroads started in 1850 and ended in 1851 and there were 243 km constructed by 1871 (Alliende 2006, pp. 14–19).

  10. 10.

    See Alliende (2006, pp. 38-72), Thompson and Angerstein (1997, pp. 76–80), Gross (1998, pp. 2–9), and Humud (1974) for more details about the construction of railroads in Chile.

  11. 11.

    Coastal navigation between cities included in our analyses was rare (Veliz 1961; Cariola and Sunkel 1990). See Summerhill (2005) and Coatsworth (1979) for a similar argument in the case of Brazil and Mexico. See McGreevey (1989) for Colombia, a Latin American case closer to the USA

  12. 12.

    This section is based on Humud (1974), Mamalakis (1976), Cariola and Sunkel (1990), Meller (1998), Badia-Miro and Ducoing (2015), and the data come from Diaz et al. (2016).

  13. 13.

    We observe convergence with respect to the USA and a period of faster growth than other countries in Latin America, with the ratio going up from 1.33 to 1.61 (the value of this ratio is 0.43 for the history of Chile). Countries in Latin America include Argentina, Brazil, Mexico, and Peru due to data restrictions.

  14. 14.

    Cariola and Sunkel (1990) documents that in 1917, 53% of the agriculture area corresponds to grains (of which 87% corresponds to wheat), 31% to forage, 8% to sylviculture, and the remainder to other crops.

  15. 15.

    Only 4% of wheat was produced in the north in 1877–1878 and about 1% in 1917–1918. The south of the Bio-Bio region corresponds to what previous research calls the Frontier (Garcia-Jimeno and Robinson 2011).

  16. 16.

    Censuses are available at the National Statistics Bureau (INE, Web page www.ine.cl). We use department-level data because it is the smallest administrative unit we can construct in panel data form. See Table 10 for details, including the area of the departments (in km2) and the populations in 1865 and 1920.

  17. 17.

    We grouped occupations to make reliable comparisons. Table 11 presents information for the 15 occupations and an occupational socioeconomic score following Duncan (1961) and Katz (2018).

  18. 18.

    Note that information for occupational variables is missing in the 1885 census.

  19. 19.

    The average department during the entire period had 26,618 people living in urban areas. The estimated coefficient of 0.225 implies an increase of 0.225×25,618 = 5975.

  20. 20.

    It is important check this for two reasons: (i) this year does not add much statistical information because railroad construction was delayed due to the Pacific War, and (ii) this is the only year for which we do not have information on occupations, so it is useful to present the correlation for later comparisons.

  21. 21.

    For instance, looking at panels (a) and (b) in Fig. 4, we note that in order to get to, for instance, the Concepcion department, railroads had to be constructed before in departments located closer to Santiago. Then, our predicted construction for Concepcion in 1865 is zero and only becomes positive in 1875.

  22. 22.

    This idea implies that the timing of the treatment and our instrumental variable is correlated with distance to Santiago, which also applies to our second instrument below. However, Table 18 shows that all results are robust to include an interaction of the treatment with distance to Santiago as control.

  23. 23.

    In particular, we consider one straight line from Santiago to Valparaiso and another from Santiago to Puerto Montt. Then, using the line from Santiago to Puerto Montt, we add another segment, defined as the shortest straight line between Concepcion and the straight line between Santiago and Puerto Montt.

  24. 24.

    As a robustness check, we also use a measure of the distance from the centroid of the department to the closest straight line and normalize the distance to make the distance to the straight line comparable to the dummy, with a value of 1 when the centroid of the department is on the straight line and a value of 0 when the distance is equal to the maximum distance observed in the data.

  25. 25.

    Results are also robust to using other distance cutoffs. Given that all results are robust to the use of spatial correlation, Table 6 is the only one reporting Conley standard errors.

  26. 26.

    Table 15 studies whether the effects of railroads are driven by the extensive or intensive margin. Using the two instruments, we construct “extensive margin” instruments and run regressions of population on our main measure of railroads penetration in a department and a dummy of whether the department has any railroads using four instrumental variables. Results suggest that the intensive margin is more important.

  27. 27.

    These ideas follow from models in Desmet and Rossi-Hansberg (2014), Rossi-Hansberg and Wright (2007), Fajgelbaum and Redding (2018), Adamopoulos (2011), Vandenbroucke (2008), Herrendorf et al. (2008), Caselli and Coleman (2001), Donaldson (2018), and Donaldson and Hornbeck (2016), among others.

  28. 28.

    Note that while our variable for agriculture expansion has a mean of 0 (as it is constructed using the residual of a regression), we demeaned the \(\log \) of the real exchange rate in the regressions; therefore, the main effect of railroads represents the impact of this variable at the average real exchange rate.

  29. 29.

    Table 17 presents results using the interaction of the real price of wheat instead of the real exchange rate. The sign and magnitude of the interactions are consistent with results in Table 8, confirming that the heterogeneous effects of railroad with the real exchange rate are related to production incentives.

  30. 30.

    Our proxy for the potential of agricultural growth is orthogonal to distance to Santiago because it is the negative of a residual of a regression including distance to Santiago. The same applies to initial population.

  31. 31.

    We also studied whether railroads affected the size of the transport and communication sector, and state presence. Any of them can be interpreted as an alternative explanation for the potential effects of railroads on population and other variables. We do not find effects of railroads on these dimensions (Table 19).


  1. Adamopoulos T (2011) Transportation costs, agricultural productivity, and cross-country income differences. Int Econ Rev 52(2):489–521

    Article  Google Scholar 

  2. Alesina A, Giuliano P, Nunn N (2013) On the origins of gender roles: women and the plough. Q J Econ 128(2):469–530

    Article  Google Scholar 

  3. Alliende MP (2006) Historia del Ferrocarril en Chile. Goethe Institut-Pehuén Editores, Santiago

    Google Scholar 

  4. Aschauer DA (1989) Is public expenditure productive? J Monet Econ 23(2):177–200

    Article  Google Scholar 

  5. Atack J, Bateman F, Margo R (2010) Did railroads induce or follow economic growth? urbanization and population growth in the american midwest, 1850-60. Social Science History 34(2):171–197

    Google Scholar 

  6. Badia-Miro M, Diaz J (2017) The impact of nitrates on the chilean economy, 1880–1930. Palgrave studies in economic history: the first export era revisited reassessing its contribution to Latin American economies

  7. Badia-Miro M, Ducoing C (2015) The long run development of chile and the natural resources curse. linkages, policy and growth, 1850-1950. In: Marc Badia-Miro VP, Willebald H (eds) Natural resources and economic growth. Learning from History. Routledge

  8. Banerjee A, Duflo E, Qian N (2020) On the road: access to transportation infrastructure and economic growth in China. Journal of Development Economics

  9. Brakman S, Frankopan P, Garretsen H, Marrewijk CV (2019) The new Silk Roads: an introduction to China’s Belt and Road Initiative. Cambridge Journal of Regions, Economy and Society

  10. Cariola C, Sunkel O (1990) Un Siglo de Historia economicá de Chile: 1830-1930. Santiago de Chile, Universitaria

    Google Scholar 

  11. Caselli F, Coleman WJ (2001) The U.S. structural transformation and regional convergence: a reinterpretation. J Polit Econ 109(3):584–616

    Article  Google Scholar 

  12. Coatsworth JH (1979) Indispensable railroads in a backward economy: the case of Mexico. J Econ Hist 39(4):939–960

    Article  Google Scholar 

  13. Conley T (1999) GMM estimation with cross sectional dependence. J Econ 92(1):1–45

    Article  Google Scholar 

  14. Desmet K, Rossi-Hansberg E (2014) Spatial development. Am Econ Rev 104(4):1211–43

    Article  Google Scholar 

  15. Díaz J, Gallego F, Lafortune J (2016) Nacimientos fuera del matrimonio en la historia de Chile: Algunos hechos estilizados. Estudios Publicos

  16. Diaz J, Luders R, Wagner G (2016) Chile 1810-2010. La Republica en Cifras. Historical Statistics. Ediciones Universidad Catolica de Chile

  17. Dinkelman T (2011) The effects of rural electrification on employment: new evidence from South Africa. Am Econ Rev 101(7):3078–3108

    Article  Google Scholar 

  18. Donaldson D (2018) Railroads of the Raj: Estimating the impact of transportation infrastructure. Am Econ Rev 108(4-5):899–934

    Article  Google Scholar 

  19. Donaldson D, Hornbeck R (2016) Railroads and american economic growth: a “market access” approach. The Quarterly Journal of Economics 131 (2):799–858

    Article  Google Scholar 

  20. Duflo E, Pande R. (2007) Dams. Quarterly Journal of Economics

  21. Duncan OD (1961) A socioeconomic index for all occupations. New York: Free Press of Glencoe

  22. Espinoza E (1897) Jeografía Descriptiva de la República de Chile. Imprenta y Encuadernación Barcelona

  23. Faber B (2012) Trade integration, market size and industrialization: evidence from China’s national trunk highway system. PhD thesis, London School of Economics

  24. Fajgelbaum P, Redding S (2018) Trade, structural transformation and development: evidence from Argentina 1869-1914. NBER Working Paper

  25. Fishlow A (1965) American railroads and the transformation of the antebellum economy. Harvard University Press, Cambridge

    Google Scholar 

  26. Fogel R (1962) A quantitative approach to the study of railroads in American economic growth: a report of some preliminary findings. J Econ Hist 22 (2):163–197

    Article  Google Scholar 

  27. Fogel R (1964) Railroads and American economic growth: essays in econometric history. The John Hopkins Press, Baltimore

    Google Scholar 

  28. Galor O, Mountford A (2008) Trading population for productivity: theory and evidence. Rev Econ Stud 75(4):1143–1179

    Article  Google Scholar 

  29. Galor O, Ozak O (2016) The agricultural origins of time preference. Am Econ Rev 106(10):3064–3103

    Article  Google Scholar 

  30. Garcia-Jimeno C, Robinson J (2011) The myth of the frontier. Understanding long-run economic growth: geography, institutions, and the knowledge economy

  31. González F (2019) Immigration and human capital: consequences of a nineteenth century settlement policy. Cliometrica

  32. Gross P (1998) Desarrollo urbano y ferrocarril del sur: 1860–1960 impacto en ciudades y pueblos de la red. Instituto de Estudios Urbanos, Pontificia Universidad Católica de Chile

  33. Herranz-Loncán A (2014) Transport technology and economic expansion: the growth contribution of railways in Latin America before 1914. Journal of Iberian and Latin American Economic History 32(1):13–45

    Article  Google Scholar 

  34. Herrendorf B, Rogerson R, Valentinyi Á (2008) Growth and structural transformation. NBER Working Paper

  35. Hummels D (2007) Transportation costs and international trade in the second era of globalization. J Econ Perspect 21(3):131–154

    Article  Google Scholar 

  36. Humud C (1974) Política económica Chilena desde 1830 a 1930. Estudios de Economiá

  37. Hurtado C (1966) Concentración de poblacion y desarrollo economico: El caso chileno. Universidad de Chile

  38. Johnson AL (1978) Internal migration in Chile to 1920: its relationship to the labor market, agricultural growth, and urbanization PhD Thesis, University of California, Davis

  39. Katz O (2018) Railroads, economic development, and the demographic transition in the United States

  40. Kujovich MY (1970) The refrigerator car and the growth of the American dressed beef industry. Business History Review 44(4):460–482

    Article  Google Scholar 

  41. Mamalakis M (1976) The growth and structure of the Chilean economy: from independence to Allende. Yale University Press, London

    Google Scholar 

  42. McGreevey WP (1989) Historia económica de Colombia 1845-1930. Tercer Mundo Editores, Bogotá

    Google Scholar 

  43. Meller P (1998) Un Siglo de Economía Política Chilena (1890-1990). Editorial Andrés Bello, Santiago

    Google Scholar 

  44. Michaels G (2008) The effect of trade on demand for skill: evidence from the interstate highway system. Review of Economics and Statistics 90 (4):683–701

    Article  Google Scholar 

  45. Miller R (1976) Social and economic change in modern Peru, chapter Railways and economic development in Central Peru 1890-1930. Center for Latin American Studies, The University of Liverpool

    Google Scholar 

  46. North D (1958) Ocean freight rates and economic development 1750-1913. J Econ Hist 18(4):537–555

    Article  Google Scholar 

  47. Nunn N, Puga D (2012) Ruggedness: the blessing of bad geography in Africa. The Review of Economics and Statistics

  48. Oppenheimer R (1976) Chilean transportation development: the railroad and socio-economic change in the Central Valley, 1840-1885 PhD thesis, University of California, Los Angeles

  49. Perez S (2018) Railroads and the rural to urban transition: evidence from 19th-century Argentina

  50. Ramírez M. T. (2001) Los ferrocarriles y su impacto sobre la economía colombiana. Journal of Iberian and Latin American Economic History 19(1):81–122

    Article  Google Scholar 

  51. Rossi-Hansberg E, Wright MLJ (2007) Establishment size dynamics in the aggregate economy. Am Econ Rev 97(5):1639–1666

    Article  Google Scholar 

  52. Rostow WW (1967) Las Etapas del Crecimiento Economicó. F.C.E

  53. Summerhill WR (2005) Big social savings in a small laggard economy: railroad-led growth in Brazil. J Econ Hist 65(1):72–102

    Article  Google Scholar 

  54. Thompson I, Angerstein D (1997) Historia del Ferrocarril en Chile. Dirección de Bibliotecas, Archivos y Museos

  55. Uribe-Castro M (2019) Caffeinated development: exports, human capital, and structural transformation in Colombia

  56. Vandenbroucke G (2008) The U.S. western expansion. Int Econ Rev 49(1):81–110

    Article  Google Scholar 

  57. Vasallo E, Matus C (1943) Historia de los Ferrocarriles en Chile. Editorial Rumbo

  58. Veliz C (1961) Historia de la marina mercante de Chile. Editorial Universitaria

  59. Verniory G (2001) Diez años en Araucanía, 1889-1899. Pehuén Editores

  60. World Bank (2007) Evaluation of World Bank support to transportation infrastructure. World Bank Publications, Washington

    Google Scholar 

  61. Zegarra LF (2011) Railroads in Peru: how important were they. Revista Desarrollo y Sociedad, 68

Download references


July 2020. We thank Oded Galor (the editor), two anonymous referees, José Díaz, Jeanne Lafortune, Rolf Lüders, José Tessada, Javier Turén, Alejandro Vicondoa, Gert Wagner, and seminar participants at PUC-Chile, SECHI meetings, and LACEA-LAMES meetings for valuable comments and suggestions. We would like to thank FONDECYT (Project 1170956) for financial support. Rodrigo Icarán, Antonia Paredes, José D. Salas, Alejandro Saenz, Felipe Vial, and Cristine Von Dessauer provided outstanding research assistance.


We received financial support from FONDECYT (Project 1170956).

Author information



Corresponding author

Correspondence to Francisco A. Gallego.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Responsible editor: Oded Galor



Table 10 Definition, departments
Table 11 Definition, occupations
Table 12 Estimation, potential for agricultural expansion
Table 13 Robustness, falsification exercises
Table 14 Robustness, main estimates
Table 15 Robustness, extensive and intensive margin
Table 16 Mechanisms, robustness to macroeconomic variables
Table 17 Mechanisms, robustness to the price of wheat
Table 18 Mechanisms, additional exercises
Table 19 Mechanisms, alternative channels

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Forero, A., Gallego, F.A., González, F. et al. Railroads, specialization, and population growth: evidence from the first globalization. J Popul Econ (2020). https://doi.org/10.1007/s00148-020-00804-3

Download citation


  • Railroads
  • Specialization
  • Agriculture
  • Population

JEL Classification

  • F14
  • N16
  • N76
  • O1
  • R1
  • R4