On the origins of ISI: The Latin American cement industry, 1900-30

Journal of Latin American Studies

Volumn 39, Issue 2, 2007, Pages 299-328

  Tafunell, Xavier ^a  

^a DEPARTMENT OF ECONOMICS AND BUSINESS   (Spain)

Author keywords

Cement industry; Economic history; Import substitution industrialisation (ISI); Latin America

Indexed keywords

CEMENT; ECONOMIC HISTORY; INDUSTRIAL DEVELOPMENT; INDUSTRIAL PRODUCTION; INDUSTRIALIZATION;

LATIN AMERICA;

EID: 34248572742     PISSN: 0022216X     EISSN: 1469767X     Source Type: Journal    
DOI: 10.1017/S0022216X07002398     Document Type: Article

References (98)

1. Stephen H. Haber, 'Introduction: Economic Growth and Latin American Historiography', in Stephen H. Haber (ed.), How Latin American fell behind (Stanford, 1989), pp. 1-33.
2. Cited from the Spanish translation of the updated version, Colin M. Lewis, 'La industria en América Latina antes de 1930', in Tulio Halperín Donghi et al., Historia económica de América Latina desde la independencia a nuestros días (Barcelona, 1997), p. 202.
3. For historical work on domestic industrialisation processes, see the review of the bibliography by Stephen H. Haber, 'The Political Economy of Industrialization', in Victor Bulmer-Thomas, John H. Coatsworth and Roberto Cortés Conde (eds.), Cambridge Economic History of Latin America. Vol II: The Long Twentieth Century (Cambridge, 2006), pp. 717-22.
4. Christian Anglade and Carlos Fortín, The State and Capital Accumulation in Latin America (Basingstoke, 1985).
5. Until the 1980s the argument that the origins of industrialisation were to be found well before 1930 generated more debate than new empirical evidence. However, historical work in the last two decades has placed more emphasis on the latter: See, for example, the work on the Argentine case by Juan C. Korol and Hilda Sabato, 'Incomplete Industrialization: An Argentine Obsession', Latin American Research Review, vol. 25, no. 1 (1990), pp. 7-30,
6. and Fernando Rocchi, Chimneys in the Desert: Industrialization in Argentina During the Export Boom Years, 1870-1930 (Stanford, 2006).
7. Enrique Cárdenas, José Antonio Ocampo and Rosemary Thorp (eds.), An Economic History of Twentieth Century Latin America (Basingstoke, 2000), vol. I, p. 23
8. Enrique Cárdenas, José Antonio Ocampo and Rosemary Thorp (eds.), An Economic History of Twentieth Century Latin America (Basingstoke, 2000), vol. III, p. 2
9. José Antonio Ocampo, 'América Latina y la economía mundial en el siglo XX largo', El Trimestre Económico, Vol. 71, no. 284 (2004), pp. 725-88
10. Haber, The Political Economy, pp. 537-40.
11. For a more detailed assessment, see Victor Bulmer-Thomas, The Economic History of Latin America Since Independence (Cambridge, 2003), pp. 127-36.
12. William R. Summerhill, 'The Development of Infrastructure', in Victor Bulmer-Thomas, John H. Coatsworth and Roberto Cortés Conde (eds.), Cambridge Economic History of Latin America (Cambridge, 2006), Vol. II, pp. 293-326.
13. Haber, The Political Economy, passim.
14. John H. Coatsworth and Alan M. Taylor, 'Introduction', in J. H. Coatsworth and A. M. Taylor (eds.), Latin America and the World Economy Since 1800 (Cambridge, 1998), pp. 1-17.
15. Concrete is a wet mixture of Portland cement, sand and aggregate. Strengthened with steel rods, 'reinforced concrete' is a substitute for steel. In Mexico around 1910 it began to be used in public works, but it was still rarely found in buildings in 1920: See Federico Sánchez Fogarty, Medio siglo de cemento en México (Mexico City, 1957), p. 13.
16. However, by this time it was considered the best material for the structures of buildings in Argentina and Uruguay: See W. W. Ewing, Construction Materials and Machinery in Argentina and Bolivia (Washington, 1920), pp. 64 and 76-7
17. W. W. Ewing, Construction Materials and Machinery in Uruguay (Washington, 1920), pp. 16-7.
18. In the two Southern Cone countries it was used in all public works. Brazil was a little slower to introduce the material, due to its great economic disparities. Cement and concrete were, however, already widely used in construction activities carried out in big cities; see W. W. Ewing, Construction Materials and Machinery in Brazil (Washington, 1920), p. 25.
19. On the other hand brick, adobe, stone and wood were still the main construction materials used in Bolivia around 1920, cement being hardly used: See W. W. Ewing, Construction Materials and Machinery in Argentina and Bolivia, pp. 161 and 173. We can assume that this was also true for the region's poorest economies.
20. A. J. Francis, The Cement Industry 1796-1914: A history (Newton Abbot, 1977), p. 26.
21. I use the expression from E. J. Hadley, The Magic Powder: History of the Universal Atlas Cement Company and the cement industry (New York, 1945), p. 62.
22. Francis, The Cement Industry, pp. 115-40 and 231-56.
23. Hadley, Magic Powder, p. 33.
24. Once again the situation in the United States can be used as an example. According to calculations based on monthly price indices, after 1900 the price of Portland cement dropped appreciably, to such an extent that in 1912 it was two-thirds of the price that it had been in 1895 (for price series, see ). This change was in marked contrast to that of construction materials in general, which rose in price by 44 Per cent between 1895 and 1912. This implies that the price of Portland cement relative to other construction materials was halved. In the following period this tendency became more striking. Between 1918 and 1929 the price of construction materials fluctuated around two and a half times the level of 1895, while the price of Portland cement was the same.
25. For production figures by country, see Ingvar Svennilson, Growth and Stagnation in the European Economy (Geneva, 1954), pp. 282-3.
26. The world total is taken from an estimate made by Federico Federico in an article published in the Italian magazine Il Cemento, reprinted in the magazine El Cemento (Barcelona), vol. III, no. 22 (1931), pp. 92-5.
27. See Svennilson, Growth and Stagnation, pp. 282-3, for the European and North American production figures.
28. The sources for the data for the other countries and at world level are Société des Nations, Annuaire Statistique de la Société des Nations 1932/33 (Geneva, 1933), p. 131,
29. and United Nations, Statistical Yearbook 1948 (New York, 1949), p. 229.
30. According to my calculations, around 1913 about half of the cement exported by Europe outside the continent went to Latin America.
31. See Bureau of Mines, The Cement Industry of Latin America (Washington, 1940), p. 3.
32. Figures for foreign trade reveal that in most Latin American countries the actual weight of cement imports far exceeded that of other products, except for iron and steel goods and coal. Sources: Argentina: Dirección General de Estadística de la Nación, Anuario del comercio exterior de la República Argentina (Buenos Aires, 1904-31)
33. Bolivia: Dirección General de Aduanas, Comercio especial de Bolivia: Exportación-importación (La Paz, 1912-31)
34. and Oficina Nacional de Estadística Financiera, Anuario: Comercio exterior de Bolivia (La Paz, 1929-32)
35. Brazil: Directoria de Estatística Comercial (from 1929, Departamento Nacional de Estatística), Commercio exterior do Brazil (Rio de Janeiro, 1912-31)
36. Chile: Oficina Central de Estadística, Estadística comercial de la República de Chile (Valparaiso, 1901-16),
37. and Anuario estadístico de la República de Chile: Comercio exterior (ValParaiso/Santiago, 1917-31)
38. Colombia: Dirección General de Estadística, Comercio exterior de la República de Colombia (Bogotá, 1915-9)
39. and Departamento de Contraloría, Anuario estadístico: Comercio exterior (Bogotá, 1920-31)
40. Costa Rica: Dirección General de Estadística, Anuario estadístico (San José, 1911-31)
41. Cuba: Secretaría de Hacienda, Comercio exterior (Havana, 903-31)
42. Ecuador: Dirección General de Estadística, Comercio exterior de la República del Ecuador en la década 1916-1925 (Quito, 1927)
43. and Comercio exterior de la República del Ecuador en los años 1925 y 1926 (Quito, 1928)
44. El Salvador: Dirección General de Estadística, Anuario estadístico de la República de El Salvador (San Salvador, 1914-21)
45. Estadística comercial (San Salvador, 1922-31)
46. Guatemala: Ministerio de Hacienda y Crédito Público, Memoria de las labores del Ejecutivo en el ramo de Hacienda y Crédito Público (Guatemala, 1926, 1929-31)
47. Haiti: Administration of the Customs, Report of the Receivership of Customs (Washington, 1919-24)
48. Financial Adviser-General Receiver, Annual Report of the Financial Adviser-General Receiver for the Fiscal Year (Washington, 1925-28)
49. and Fiscal Representative, Annual Report of the Fiscal Representative for the Fiscal Year (Washington, 1929-32)
50. Mexico: Secretaría de Hacienda y Crédito Público, Comercio exterior y navegación (Mexico, 1902)
51. Anuario de estadística fiscal (Mexico, 1913-14)
52. Anuario del comercio exterior y navegación (Mexico, 1920-28)
53. and Departamento de Estadística Nacional, Anuario Estadístico: Comercio exterior y navegación (Mexico, 1920-22); 1926-32
54. Nicaragua: Administración de Aduanas, Memoria del Recaudador General de Aduanas y las Estadísticas del Comercio (Managua, 1921-31)
55. Paraguay: Dirección Gen eral de Estadística, El comercio exterior del Paraguay (Asunción, 1928)
56. Peru: Superintendencia General de Aduanas, Estadística especial del Peru (Callao, 1902, 1912-6; 1926-31)
57. República Dominicana: Receptoría General de Aduanas, Report of the ... fiscal period. Together with summary of commerce (Washington, 1919-31)
58. Uruguay: Dirección General de Estadística, Anuario estadístico de la República Oriental del Uruguay (Montevideo, 1901-12; 1916-31)
59. Venezuela: Ministerio de Hacienda y Crédito Público, Estadística mercantil y marítima (Caracas, 1907-31).
60. The following paragraphs draw their information from Bureau of Mines, The Cement Industry, and Juan de las Cuevas Toraya, Un siglo de cemento en Latinoamérica (Mexico, 1999). These publications contain detailed and systematic information on the companies established in the region.
61. For more information, see Juan de las Cuevas Toraya, 100 años del cemento cubano (Havana, 1995).
62. See Wilson Suzigan, Indústria brasileira: Origen e desenvolvimento (São Paulo, 2000), pp. 264-71, as well as the publications cited in the following note.
63. See W. W. Ewing, Construction Materials and Machinery in Brazil, p. 45.
64. The troubled history of the first factories on Brazilian soil might lead one to the conclusion that Brazilian cement production did not really start until 1926, with the installation of a factory in Perús (São Paulo state) by the Companhia Brasileira de Cimento Portland, as argued by the main authorities on the early stages of Brazilian industrialisation: See A. Villela y W. Suzigan, Política do governo e crescimento da economia brasileira (Rio de Janeiro, 1977)
65. and Claudio Haddad, 'Crescimento do produto real brasileiro, 1900/1947', in F. R. Versiani and J. R. Mendonçca de Barros (eds.), Formação econômica do Brasil: A experiẽncia da industrialização (São Paulo, 1977).
66. For more details see Carlos M. Peláez, História da industrialização brasileira (Rio de Janeiro, 1972), pp. 195-208. I have applied the same criteria with respect to Argentina and have not taken into account the very small amounts of cement which could have been produced before the opening of the Fábrica Nacional de Cemento Portland in 1908.
67. In fact only three were destined for a short life, the two plants opened in Cuba and the one opened in Colombia in 1913. The first plant (El Almandares, located in Havana) closed down in 1921, unable to cope with competition from a new factory installed by North Americans in Mariel in 1918. The second had an almost ephemeral existence. Built between 1912 and 1913 in the vicinity of Guantánamo, it closed down after the outbreak of the First World War. Finally, the survival of the Compañía Industrial de Cemento Antioqueño, which kept an extremely inefficient plant running in the district of Medellín between 1913 and 1919, was possible only because of the exceptional circumstances of the time. In Guatemala, close to the capital, the small factory owned by Carlos Novella (1,500 tonnes) was established and is still running today. In Mexico it seems that two local enterprises started small-scale production at the beginning of the century. They soon folded: See Sánchez Fogarty, Medio siglo de cemento, p. 6.
68. Production of Portland cement started in earnest in Mexico with the company Cementos Hidalgo, sited in the town of the same name in Nuevo León, with a capacity of 36,000 tonnes. Three years later it was joined by La Tolteca, located on the outskirts of Mexico City. The Cruz Azul company commenced operations in 1910. All three companies are still active today. In 1908, Argentina and Chile both followed by establishing factories. The Fábrica Nacional de Cemento Portland opened in Rodríguez del Busto (province of Córdoba, Argentina) with a production capacity of 12,000 tonnes, while in Chile the Fábrica de Cemento El Melón, located in La Calera, halfway between Valparaiso and Santiago. opened with an initial potential capacity of 40,000 tonnes, which made it for several years the largest producer of cement in Latin America: See Gabriel Palma, 'From an Export-Led to an Import-Substituting Economy: Chile, 1914-39', in Rosemary Thorp (ed.), Latin America in the 1930s: The Role of the Periphery in World Crisis (Basingstoke, 2000), pp. 44-5. The Argentine company lasted twenty years, while the Chilean is still going strong. In 1909 Colombia and Venezuela joined the cement producing countries, the former with a plant created by the Samper family close to Bogotá (with an initial Capacity of 3,600 tonnes), and the latter with the La Vega plant on the outskirts of Caracas, opened by the Fábrica Nacional de Cementos, with a capacity of 7,500 tonnes. The Samper Bush brothers' plant located at the mine at La Calera soon became obsolete and was replaced by another nearby. The Venezuelan plant stayed open until 1990. Finally, Uruguay opened a factory just when war broke out in Europe. The Fábrica Uruguaya de Portland in Sayago, on the edge of Montevideo, with a capacity of 45,000 tonnes, was in operation until 1994.
69. The six factories opened in the country before 1929 were all situated in the Mexican altiplano or high central plains, separated from the coast by mountainous terrain, making the transport of cement to the area economically impractical. This same factor was also decisive in the establishment of the cement industry in Bolivia, although the country's extreme poverty delayed this until 1926. In Mexico's case the initial problems were caused by the civil war: See Stephen H. Haber, Industry and Underdevelopment: The Industrialization of Mexico, 1890-1940 (Stanford, 1989), pp. 126-36.
70. Foreign capital only played a significant role in the ventures in Chile (mixed Anglo-Chilean capital), in Cuba at El Almendares (French, Spanish, and Cuban capital), two of the Mexican ventures (Anglo-North American capital), and in Uruguay (originally German capital).
71. Bureau of Mines, The Cement Industry, p. 2.
72. The Compañía Cubana de Cemento Portland, located in Mariel (province of Pinar del Río), lit its first two kilns, with a capacity of some 137,000 tonnes, in 1918. This was followed a year later by the plant started in Sierras Bayas (Province of Buenos Aires) by the Compañía Argentina de Cemento Portland, with a capacity of 150,000 tonnes. Peru's turn came when the Compañía Peruana de Cemento Portland El Sol opened a small factory capable of producing 3,000 tonnes near Lima. In Ecuador the Compañía Anónima de Industrias y Construcciones installed a factory with an annual capacity of 20,000 tonnes in Estero Salado, close to Guayaquil. Mexico saw its productive base broaden in 1923 with the firm Cementos Portland Monterrey, in the state of Nuevo León (45,000 tonnes capacity) and the Compañía de Cemento Portland Landa (3,000 tonnes) in Puebla.
73. The factory in Olavarría (Province of Buenos Aires), owned by the Compañía Industrial Argentina Loma Negra, added 80,000 tonnes to the nation's potential output. A year later, Juan Minetti Canteras added a further 120,000 tonnes with his plant in the Dumesnil station, in the Province of Córdoba.
74. As noted already, the Companhia Brasileira de Cimento Portland, opened a factory in Perús (São Paulo state), with an initial capacity of 60,000 tonnes. In Viacha, close to La Paz, the Sociedad Boliviana de Cemento built a small factory with the capacity to produce 2,000 tonnes.
75. In truth, the Paraguayan initiative came about in 1912, sponsored by a French investment group, but failed as a consequence of the war. In 1926 a local company managed to repair and finish work on the factory but was not able to run it on a regular basis.
76. Lone Star carried out its investment policy through its financial subsidiary, the International Cement Corporation. This corporation, with its headquarters in New York, created the Compañía Cubana de Cemento Portland, the Compañía Argentina de Cemento Portland, and the Compañía Peruana de Cemento Portland. It also acquired the Uruguayan cement factory which it completely renovated and registered as Compañía Uruguaya de Cemento Portland. The crisis of the 1930s was no check to Lone Star, which went on to broaden its portfolio of investments. Independently, Canadian and British capital provided the bulk of the funds for the Companhia Brasileira de Cimento Portland. Peruvian and North American investors put up the money for the Bolivian cement business. In Ecuador, German capital played an important part, conditional on the factory being equipped by Krupp. As for the small factory set up in 1901 by the engineer Novella in Guatemala, equipped with second-hand German machinery, it was modernised in 1917 thanks to an agreement with United Fruit, which provided the majority of the capital. Finally, in 1923 an Anglo-Canadian consortium took over the factory in Ecuador.
77. During the first wave of development the German companies, Polysius and Krupp, and the Danish company, F. L. Smidth, were among the major suppliers of technology used in Latin American installations. After the second wave of development the cement industry was fitted out with machinery which came, in approximately equal parts, from the United States and Europe. See Kock-Petersen, 'The Cement Industry', in L. J. Hughlett, Industrialization of Latin America (New York, 1946), p. 69.
78. At the start of the Second World War, according to my calculations, the total production of the Latin American cement industry already covered 86 per cent of demand (apparent consumption). After the war, this high level was not reached again until 1953. Following this, the process of ISI was quickly completed.
79. Detailed and comprehensive information is available on how the production capacity of the cement industry in all the Latin American and Caribbean countries developed, thanks to the research work of De las Cuevas, Un siglo de cemento. Where information is not available, two hypothetical production values have been calculated, based on the capacity of the country's installations, one corresponding to a maximum probable level (90 per cent of installed capacity) and the other to a minimum probable level (20 per cent of installed capacity). These percentages are not arbitrary but appear to be the most realistic after examining the known empirical evidence. It was only under exceptional circumstances that cement factories went beyond die upper limit or reduced their production to below the minimum numbers here; the first was difficult to exceed for technical reasons, while to go below the second would have been unsustainable for reasons of economic profitability.
80. This percentage should be taken only as an estimate, since it is an abstraction of the margins of error existent in the production figures. These margins vary between ± 4 or 5 percentage points.
81. 2 figures are 0.633 and 0.449.
82. Cuba may serve as an example of the extraordinary importance that, on certain occasions, public works could have. See 'Evolución y perspectivas de la industria del cemento en Cuba', Revista del Banco Nacional de Cuba vol. 2, no. 7 (July 1956), p. 29.
83. This case also highlights the fact that, on occasion, public sector demand could consolidate the development of the domestic cement industry. The special programme of public works approved in 1926 by the Cuban government contained an agreement guaranteeing supply at a prearranged price with the company which had a virtual monopoly on domestic production. This agreement meant that the company was able to work at full output. See De las Cuevas, 100 años de cemento, pp. 35-6. Unfortunately, historians of Latin America have yet to draw up figures for public investment for this period which cover the whole region.
84. Kock-Petersen, The Cement Industry.
85. As has been shown, the Latin American cement industry depended on the combination of technology and investment provided by the most industrialised economies. It seems reasonable to assume that during this period there was little difference in the receptiveness of the Latin American republics to these initiatives. As for the supplies required for cement production, a distinction has to be made between energy resources and the main raw materials (limestone and clay). The kilns were fuelled by coal or oil. Although fuel consumption was very high, the common assumption that the cement industry was extremely energy-intensive has to be qualified. Taking the Spanish industry as an example, between 1922 and 1928 it used a quantity of coal equivalent to 31 per cent of the cement produced (both expressed in tonnes). We have to remember that the proportion of coal, as weight, in relation to mineral raw materials, would only have been around 17 per cent, given that almost two tonnes of raw materials produced one tonne of cement. See the cement production figures and coal consumption figures in Cemento (Barcelona), vol. I, no. 5 (1929), pp. 159-60.
86. Suzigan, Indústria brasileira, pp. 264 and 267.
87. Tariff policy possibly inhibited the development of a national cement industry in many countries, while in a minority it became a powerful stimulant. This is an aspect of the subject that is greatly in need of research. In Porfirian Mexico, thanks to the work of Graciela Márquez ('Tariff Protection in Mexico, 1892-1909: Ad Valorem Tariff Rates and Sources of Variation', in Coatsworth and Taylor (eds.), Latin America pp. 407-42), it is known that the government clearly encouraged the sector in 1905 by strengthening the protectionist policy already in place. On the other hand, numerous countries maintained a very liberal policy, further skewed by the fiscal privileges conceded both to the public sector and to the large utilities companies (railways, electric companies). This policy was governed by the principle that construction materials could be imported free of import duties by the public sector and these utilities companies while other users had to pay more or less moderate tariffs. Most of the small Central American and Caribbean nations, realising they had practically zero possibilities of starting a domestic industry, inclined towards a liberal policy of economic encouragement; cement was imported without the payment of duties, in common with other supposedly strategic goods such as machinery.
88. This fact should be emphasised, not only because it is little known, but also because of recent arguments that Latin American states were the most protectionist in the world: See John H. Coatsworth and Jeffrey G. Williamson, 'Always protectionist? Latin American Tariffs from Independence to Great Depression', Journal of Latin American Studies vol. 36, no. 2 (2004), pp. 205-32.
89. Fernando Rocchi, Chimneys in the Desert, p. 94.
90. If the average level of the first three years (1900-1902) is compared with the last three (1928-30), the per capita consumption multiplies by eight.
91. If we accept, as a reasonable approximation, the information compiled by Angus Maddison, The World Economy: Historical Statistics (Paris, 2003), in 1900 per capita GDP in the richest country (Argentina) was four times higher than the poorest (Brazil) within the group of the eight large economies (there being no data for the rest). In 1929 the gap between Argentina and Brazil had reduced very slightly. It must be remembered that some small Central American and Caribbean countries, as well as maybe Bolivia and Paraguay, very possibly had a per capita income below that of Brazil.
92. For the consumption of modern energy sources see César Yáñez, María M. Rubio and Albert Carreras, 'Economic Modernisation in Latin America and the Caribbean between 1890 and 1925: A View from the Energy Consumption', 44th Cliometrics Conference, June 2-4, 2006 (Binghamton, New York). The ranking for the consumption of cement per inhabitant in the Latin American nations corresponds exactly to the ranking for the consumption of energy.
93. If we assume that production was equivalent to consumption in the case of the United States, Latin American per capita levels of consumption would be, according to my calculations, based on information from Svennilson, Growth and Stagnation, 13.5 per cent of those of the powerful northern neighbour in both 1913 and 1925.
94. Ocampo,'América Latina'.
95. Haber, Industry and Underdevelopment, pp. 32 and 40, attributes the status of a natural monopoly to the cement industry of the time, due to the fact that high transportation costs allowed the product to be sold only within a radius of 250 kilometres. But the exceptional nature of the Mexican case should not be overlooked: The Mexican industry enjoyed natural protection for geographical reasons. Its counterparts in the remainder of Latin America were not so fortunate.
96. This seems to contradict the thesis which states that the war did not favour industrial development in the region, even in the places where it was most advanced: See Rory Miller, 'Latin American Manufacturing and the First World War: An Exploratory Essay', World Development, vol. 9, no. 8 (1981), pp. 707-16,
97. and Bill Albert, South America and the First World War: The Impact of the War on Brazil, Argentina, Peru, and Chile (Cambridge, 1988). This thesis is probably correct in overall terms, but not when referring to the cement industry, given its specific characteristics.
98. The spectrum is drastically reduced if we compare the second consumer with the second-last (Argentina and Colombia, respectively), although the ratio of 118:1 which found between them bears no proportion to their differential in terms of per capita income, which, according to figures from Maddison, would not be more than 3:1. See Maddison, The World Economy, table 4c.