Assessing The Obstacles To Industrialisation: The Mexican Economy, 1830-1940

Assessing the Obstacles to Industrialisation: The Mexican Economy, 1830-1940 Author(s): Stephen H. Haber Source: Journal of Latin American Studies, Vol. 24, No. 1 (Feb., 1992), pp. 1-32 Published by: Cambridge University Press Stable URL: http://www.jstor.org/stable/156670 Accessed: 21/04/2009 15:57 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=cup. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access to Journal of Latin American Studies.

http://www.jstor.org

Assessing the Obstacles to Industrialisation: The Mexican Economy, STEPHEN

83 0-I94o*

H. HABER

After England began what came to be known as the First Industrial Revolution at the end of the eighteenth century, industrial technology quickly diffused throughout the nations of the North Atlantic. Within fifty years of the first rumblings of British industrialisation, the factory system had spread to Western Europe and the United States. Latin America, however, lagged behind. It was not until the twentieth century that manufacturing came to lead the economies of Latin America and that agrarian societies were transformed into industrial societies. This article seeks to understand this long lag in Latin American industrialisation through an analysis of the experience of Mexico during the period 8 30-1940. The purpose of the paper is to look at the obstacles that prevented self-sustaining industrialisation from taking place in Mexico, as well as to assess the results of the industrialisation that did occur. The basic argument advanced is that two different types of constraints prevailed during different periods of Mexico's industrialisation. During the period from 1830 to 1880 the obstacles to industrialisation were largely external to firms: insecure property rights, low per capita income growth resulting from pre-capitalist agricultural organisation, and the lack of a national market (caused by inefficient transport, banditry and internal tariffs) all served as a brake on Mexico's industrial development. During * Earlier versions of this article were presented at the 'Why the Lag in Latin American Industrialization?' conference at Harvard University, the Stanford University Social Science History Workshop, and the Conference of the California Intercampus Group in Economic History. Jeffrey L. Bortz, Frederick P. Bowser, Gregory Clark, John H. Coatsworth, Kenneth L. Sokoloff and John D. Wirth, as well as an anonymous referee, made detailed comments on earlier drafts. Research for this article was funded by grants from the Latin American Program of the Social Science Research Council and the Fulbright Program. The usual disclaimers apply. Stephen H. Haber is Associate Professor of History at Stanford University.

J. Lat. Amer. Stud. 24, I-32

Printed in Great Britain

2

StephenH. Haber

the period I880-1910 the obstacles to industrialisation were largely internal to firms. These factors included the inability to realise scale economies, high fixed capital costs and low labour productivity. During the period from 910o to 1930 these internal constraints combined with new

external

constraints -including

the Revolution

of 1910-17,

the

political uncertainty of the post-revolutionary period and the onset of the Great Depression - which further slowed the rate of industrial growth. This article is organised into three sections, each of which takes up one of these periods in detail. Each section begins with a discussion of the obstacles to industrial growth and then assesses their impact on the structure and nature of manufacturing. The article ends with the rekindling of Mexican industrial growth in the mid-193os. The outcomes of that wave of growth, which stretched through the mid- 970s, are not discussed in detail here, since that period has been written about extensively by other scholars. Mexico is chosen for study because, of all the countries of Latin America, it was the most successful industrialiser prior to the Second World War. As early as the I83os cotton textile manufacturing in Mexico began to take place in a factory setting and by the I89os Mexico was producing a wide range of goods in large-scale, vertically integrated enterprises. In fact, many of Mexico's present-day industrial giants, some of which form the nuclei of large conglomerates, date from the end of the nineteenth century. Examples can be found in the steel, cement, glass, explosive, cigarette, soap, cotton and wool textile, paper and beer industries. Yet while industrial development began in Mexico quite early relative to the rest of Latin America, the process of self-sustaining industrialisation that took hold in England or the United States did not occur. Instead, a series of obstacles produced a truncated process of industrialisation dependent on government intervention, lacking in backward linkages to capital goods production, characterised by oligopoly and monopoly production, and dominated by entrepreneurs following rent-seeking strategies. Assessing the obstaclesto industrialisation,I8No-80 In looking at the economic history of Western Europe and the United States, scholars have pointed to a number of features of those economies which encouraged rapid rates of economic growth. Among the features commonly mentioned are the availability of low-cost transport, a relatively even distribution of income, highly productive agricultural sectors, low entry costs for early manufacturers, and the development of legal institutions designed to protect property rights and facilitate commerce. In short, it is generally agreed that those economies which were already

Industrialisationin Mexico

3

affluent, which possessed large and well-integrated markets, and which had the ability to accumulate and mobilise the capital to finance manufacturing ventures were the economies which were the first to industrialise. Scholars have emphasised that no single feature or institution can be isolated as the determinant in industrial success, but they would agree that the lack of these features as a group would certainly work against successful industrialisation. It just so happens that from the i830s to the I88os almost all of the features which permitted industrial growth to occur rapidly in the Western European and US settings were lacking in Mexico. First and foremost, those countries which underwent successful processes of industrialisation were already relatively affluent. They had realised slow rates of growth in per capita income for a number of decades well before they began to industrialise, and attained quite high levels of national income for pre-industrial societies. The Mexican economy, on the other hand, was not only significantly smaller than the successful industrialisers, it was also shrinking - both absolutely and relatively. In I8oo, according to John Coatsworth's estimates, Mexico's per capita GDP (in 1950 US dollars) was $73, roughly two-fifths that of the United

States and Great Britain. Throughout the first half of the nineteenth century, per capita GDP actually declined, falling to $5 6 per capita in 845 and $49 in I86o. In this last year it was only 13 % of British per capita GDP and 14 % of the US figure. It was not until the i86os that per capita GDP began to increase, rising to $62 in 1877 and to $9' in i895.' Underlying this fall in national income were a number of factors, all of which we may also specify as obstacles to industrialisation. It should be kept in mind, however, that these were impediments not only to industrial growth, but to economic growth in general. Let us take up each of these in detail. Perhaps the most noticeable obstacle to Mexican industrialisation was the lack of a national market because of high transport costs. Unlike the United States, England or the other early industrialisers, whose regions were connected by low-cost transport systems (canal and riverine traffic, coastal shipping, and toll roads), Mexico's topography prevented the development of inexpensive modes of transport. Only one-third of the land mass of the country is reasonably level, and there aie almost no navigable waterways, save for a few rivers in the sparsely populated southeastern states of Veracruz and Tabasco. In addition, the great majority of the population has historically lived in the mountainous interior of the country, meaning that coastal traffic could not play a significant role in linking markets. 1John H. Coatsworth,'Obstaclesto Economic Growth in NineteenthCenturyMexico', American Historical Review, vol. 83 (1978), p. 82.

StephenH. Haber

4

Almost all traffic therefore had to move over mountainous terrain by expensive mule train or ox-drawn, wheeled vehicle. Transport by such means was costly because it required an expensive highway system to be constructed and maintained. It was also unreliable, as Mexico's rainy season, which stretches from May to September, regularly made the roadbeds impassable. During the colonial period a road system was indeed constructed in order to accommodate the heavy, two-wheeled carrosthat were used to transport silver from the mines in the Bajio and imported manufactures from the port city of Veracruz to Mexico City, but beginning in the eighteenth century this system fell into disrepair. After independence in I821 the situation became even worse, as there was little money available for the improvement or maintenance of the rapidly deteriorating roads. By the nineteenth century only a single roadway existed that was suitable for wheeled traffic over its entire length. Even on this highway, which ran from Mexico City to Chihuahua via Zacatecas and Durango, mule trains outnumbered wheeled traffic.2 Railroads, which would have alleviated the transport problem, were not constructed for the same reasons that affected highway maintenance. Federal funds were nowhere near sufficient to launch a railroad programme. Foreign investment could have been used to build a rail system, but none of the regimes from I82I to I876 had much luck in enticing foreign capitalists into financing such a venture. Mexico's first rail line, from Mexico City to Veracruz, was not completed until 873. As late as 1877 the entire Mexican rail grid contained only 640 kilometres of track, of which I I4 employed mules rather than steam engines as the source of motive power.3 The result of this situation was extremely high transport costs. Two examples can give the reader an idea of how difficult the transport situation was. In 82 5 the Real del Monte mining company imported from England steam engines and other equipment necessary to renovate its mines in the state of Hidalgo. The 3 so-mile journey from the port to the mines (Veracruz to Pachuca) took almost twelve months. Similar delays occurred in the transport of textile machinery. In fact, imported textile machinery often doubled in cost by the time it reached Mexico City from Veracruz, a distance of only some 300 miles.4 The transport problem was compounded by widespread banditry. By mid-century brigandage was so commonplace that the National Congress could not meet because the representatives feared travelling from their 2

John H. Coatsworth, Growth Against Development: The EconomicImpact of Railroads in 3 Ibid., p. 35. Porfirian Mexico (Dekalb, Illinois, 1981), p. 18. on Mexico's Nineteenth 'Some Issues Century Depression', mimeo 4Enrique Cardenas, (1981),

pp. 29-30.

Industrialisationin Mexico

5

home districts to Mexico City on the nation's bandit-infested roads. Though no global data are available on the frequency of highway robberies, Paul Vanderwood's qualitative data indicate that, like contemporary joggers in Central Park, most travellers expected to be robbed at least once on a long journey. The available quantitative data mirror this judgement. In one single day in 1861, for example, the stagecoach from Mexico City to Puebla was robbed three times.5 During one Is-day period in I865 there were four major robberies along the highway from Orizaba, Veracruz, to Mexico City- an average of one every four days. The Mexico City Independencia,summing up this situation,

proclaimed that it was 'a scandal the way bandits infest the roads. What the government is thinking about we do not know, but the bandits rampage with impunity.'6 In addition to inefficient transportation technology and the threat of banditry, Mexico's market was further segmented by a system of internal tariffs, similar to that which existed in Germany prior to the Zollverein. Until the latter part of the nineteenth century, Mexico was a congery of individual provinces, each of which functioned with a great degree of autonomy from the central government. Essential to the political autonomy of these provinces was the fiscal autonomy provided by state taxes. That this taxation system served as an impediment to the development of a national market and a national economy was not deemed particularly important by those who benefited from it. Since the scale of technology employed in most manufacturing enterprises in the early nineteenth century was very modest, robust regional markets probably could have sustained an industrialisation process, even without access to a national market. Indeed, as Sidney Pollard has demonstrated, industrialisation in nineteenth-century Europe was largely a regional phenomenon.7 The market for manufactured goods in mid-nineteenth century Mexico, however, was neither robust nor deep. Our knowledge of living standards, the distribution of income, and wages during the early republic is extremely limited, but indirect evidence about Mexico's social and economic structures suggests very low income levels. Indeed, the great majority of the Mexican population were villagedwelling peasants who practised rain-fed, subsistence agriculture. A sizeable portion of the population did not even function in the money economy. Even as late as 191o, after a 40-year period of economic growth and structural change, 71.3 % of the population

7

resided in villages of less

Paul Vanderwood, Disorder and Progress: Bandits, Police, and Mexican Development 6 Ibid., p. II. (Lincoln, Nebraska, I98I), pp. 3, ii and 37. Sidney Pollard, Peaceful Conquest: The Industrialization of Europe, I76o0-970 (London, 1981).

6

StephenH. Haber

than 2,500 persons, indicating that they were peasants who lived in traditional villages.8 A good portion of this population (close to 6o % in some states) could not even speak Spanish, implying that they functioned only in local markets.9 Finally, Coatsworth's national income estimates for the nineteenth century suggest an economy characterised by low income levels. Gross Domestic

Product per capita in I845 was only $56 (195o

dollars), and was, as has previously been mentioned, shrinking.'0 The market for manufactures would in fact have been even smaller than this $56 figure indicates, since the nation's skewed distribution of wealth no doubt produced an extremely uneven distribution of income. In fact, at the turn of the century, when real GDP per capita had doubled since 845, average household incomes were still so low that in years of bad corn harvests, when the price of this important staple rose, Mexico's working class could not afford to purchase cotton cloth.1" Not only was the market not conducive to the growth of manufacturing, neither was the legal-institutional environment. In the United States and England, for example, legal systems had developed in which decisions were based on precedent, in which judges made rulings based on pride in craft, and in which well-defined bodies of contract, commercial and patent laws were established. This made property rights, at least for the 'respectable classes', easy to protect. In Mexico no such state of affairs existed. Until the last two decades of the nineteenth century no legislation existed to encourage the formation of limited liability corporations, a modern patent law did not exist, nor was there a body of mortgage credit law designed to protect long-term investments. Indeed, as was the case in early modern Europe, economic activity of all kinds still required special permits and licences for which special taxes and fees were charged. Above all, litigants faced a corrupt and whimsical judicial system. Thus it was extremely difficult to enforce contracts, collect on loans (especially if they were to the government) or enforce property rights without resorting to political machinations. As David Walker has pointed out in his work on commerce in nineteenth-century Mexico, without access to those who 8

James W. Wilkie, The Mexican Revolution:Federal Expenditureand Social Changesince o191 (Berkeley,

1970), p. 2I8.

9 Direcci6n General de Estadistica, Estadisticas Sociales del Porfiriato, r877-y9ro (Mexico, 1956), pp. II8-20.

The data are for I895, when in the nation as a whole

17% of the

population spoke only an indigenous language. The percentage of non-Spanish speakers in the mid-nineteenth century was certainly significantly higher than the 1895 figure. But even in I895 the majority of the population in indian-dominated states like Yucatan and Oaxaca spoke only an indigenous language. In seven other states (out of 3 ) over 20 % of the population spoke no Spanish in I895. It is highly likely that in the mid-nineteenth century the national average approached the levels of these states. 10 Coatsworth, 'Obstacles to Economic Growth', p. 82. 1 El Economista Mexicano, 7 May I904, p. 114.

Industrialisationin Mexico

7

wielded political power and influence it was virtually impossible to conduct business of any kind.12 Since the government changed hands in the early nineteenth century on an almost monthly basis (Mexico had 75 presidents between i821 and I876), access to those who wielded the political power necessary to enforce property rights constantly shifted as well. The net result was an arbitrary institutional environment that made free enterprise problematic. The ramifications of these various institutional and social factors were several. First, it meant that the Mexican economy could not grow. Secondly, because of the lack of a robust economy, the Mexican state could not defend itself from external or internal attack. Indeed, it was during this period that the nation lost nearly half of its territory to the United States, saw the secession first of the Central American provinces and later the Yucatan, and was successfully invaded by France. Finally, these factors served to discourage industrial development. Manufacturers faced a divided and shallow market, an uncertain institutional environment and a contracting economy. What is perhaps most surprising about the economic history of this period is the amount of industrial development that did occur. The history of early Mexican manufacturing is still far from written, but the picture that does emerge indicates an impressive amount of industry relative to the other countries of Latin America. In a way, some of this industrialisation was due to the very features that discouraged a more complete and self-reinforcing process of industrial growth. High internal transport costs protected small manufacturers from the onslaught of British goods. The ability of the well-connected to manipulate the state to their own advantage produced a government finance bank (capitalised from import revenues) which lent the majority of its portfolio to its own board members. The need for government revenues produced tariffs, both federal and state, which served to afford some protection to industry. The best-developed and fastest-growing of Mexico's nineteenth-century industries was cotton textiles. In 1843, after a I2-year period of investment and growth (a small fraction of which was subsidised by loans from the government-sponsored

Banco de Avio, which existed from 1830 to 1842),

the nation boasted 59 cotton textile factories, the great majority of which were located in the states of Mexico and Puebla, producing primarily for the Mexico City market. These firms processed some io.6 million kilos of raw cotton into yarn, most of which was sold to independent weavers though some was also processed into cloth in the factories. No data are 12

Coatsworth, 'Obstacles to Economic Growth', p. 98; David W. Walker, Business, Kinship and Politics: The Martinet de Rio Family in Mexico, I82j-1867 (Austin, Texas, 1986).

8

StephenH. Haber

available on the number of workers employed in I 842, but it was probably in the area of ten thousand.13 Eleven years later the number of active mills had actually declined to only 42. Other indicators suggest, however, that the situation had improved for the remaining mills. If we use raw cotton consumption as a proxy for output (which eliminates the problems associated with measuring output in goods of different types and different qualities over time), the data indicate a rise in production

of I9 %.14 This increase may

have been due to the increase in the number of machines in service, which grew by I8 %. Lack of accurate labour-force data makes it difficult to assess the growth of labour productivity. By 877, the end of the early period of Mexican manufacturing, the size of the cotton-textile industry had increased significantly since the I85os. Mexico now had 92 cotton mills, though these were somewhat smaller in size than the I854 factories: the average mill now worked only 2,753 spindles with 128 workers, compared to 3,004 spindles with 264 workers

in I854.15 In comparison to the United States, this was a modest textile industry. Mexico's compared

ring spindles in 1877 was an extremely low figure to the 10.7 million spindles in the United States in I880.16

253,270

Compared to other Latin American nations, however, this was quite a sizeable industry. Brazil, for example, in 1866 possessed but nine mills, employing 768 workers and 14,875 spindles. By i88i the number of Brazilian mills had increased to 44, and now employed 60,419 spindles, but this was still only one-quarter of the number in Mexico.17 Besides cotton textiles other important industries existed as well, though none functioned on as large a scale. Unfortunately, little in the way 13

Secretariade Hacienday CreditoPLblico,Documentos parael estudiodela industrialiiacion en Mexico, I837-i84I

(Mexico, 1977), doc. 5; Ministerio de Fomento, Estadistica del

de Mexico(Mexico, I854), doc. 2. For a discussion of the Banco de Avio Departamento in the Early andIndustrialDevelopment see Robert A. Potash, The MexicanGovernment Republic:TheBancodeAvio (Amherst, I983). The literaturehas tended to overestimate the role of the bank in financingMexico'searlyindustrialisation.The total loans of the bank to cotton textile manufacturers from 1830 to 1842 came to 509,000 pesos, but in

854, when therewere fewer mills in operationthanin 1842,the total value of the plant, equipment, and buildings of the industry (valued at acquisition cost) was 8,872,95i pesos, indicatingthat the bank could only have been responsiblefor six per cent of the financecapital of these enterprises. 14 Ministerio de Fomento, Estadisticadel Departamento de Mexico,doc. 2. 5 Calculatedfrom ibid.; and Secretariade Hacienda,Estadisticade la ReptiblicaMexicana (Mexico, i88o), cuadros de industria. 16 17

US Bureau of the Census, Censusof Manufactures, 1879. A. V. de BorjaCastro,'Relatorio do segundo grupo', in Antonio Jose de Souza Rego, de 1866(Rio de Janeiro, 869), p. 49; Ministeriode Relatorioda segunda exposifaonacional Executiva Textil, Indistriatextil algodeira Comisio e Trabalho, Industria, Comercio, (Rio de Janeiro, 1946), p. 5I.

Industrialisationin Mexico

9

of time-series data is available for these other industries, but qualitative data indicate that their rates of growth were quite slow. Many resembled artisanal shops more than modern factories. In the iron and steel industry, for example, an 853 census listed but five foundries, only one of which employed more than oo00 workers. Median foundry size was but 14 employees. A similar situation prevailed in the glass industry, where an 1857 census turned up only five firms, four of which employed fewer than Ioo workers. Median firm size was 70 workers.18 Small-scale industries existed as well in paper manufacturing, beer brewing, and other activities, but again we know little about their development over time. It is clearly the case, however, that even as late as the i88os these industries were very modest in scale. In the paper industry, for example, until I892 the combined value of production of the nation's 12 paper mills was scarcely one million pesos (at that time roughly $700,000). Total output was approximately four tons of paper per day.19 In short, with the exception of cotton textiles, most of Mexican industry was still small in scale by the i88os. An 1877 census of the manufacturing establishments in the Federal District, for example, turned up 728 manufactories. Industries like boots and shoes, other leather working, hats, tailor shops, bakeries and wood-working predominated. Average firm size was only 17 workers, implying that these were more artisanal shops than modern factories. Even in industries which in the advanced economies were becoming large-scale, like iron and steel, smallscale factories continued to predominate. In the iron industry, for example, the 1877 Federal District census turned up only two foundries, whose combined investment in fixed capital was only 54,000 pesos (at this point the dollar and peso traded roughly one for one) and which employed only Ioo workers between the two of them. Total output was approximately 5oo,ooo kilos of iron goods, with a value of only 6o,ooo pesos.20

Most significantly, all of the firms enumerated in the 1877 census were consumer goods producers; no capital or intermediate goods industries existed. Indeed, throughout the nation this was the case. The explanation for this situation is fairly straightforward: the consumer goods industries were still too underdeveloped to support capital goods industries. Even in the relatively rapidly growing cotton textile industry, an industrial 18

Secretaria del Estado, Memoria de la Secretaria del Estado y del Despacho de Fomento ColonigacionIndustriay Comerciode la Repiblica Mexicana, ir87 (Mexico, I857), docs. 182, I8-3.

19 Hans Lenz and Federico G6mez de Orozco, La industriapapelera en Mexico (Mexico, 1940),

p. 83.

20 Secretaria de Hacienda, Estadzsticade la ReptiblicaMexicana (Mexico, I880), cuadros de

industria.

o

StephenH. Haber

plant consisting

of only 92 small factories

running

just over 250,000

spindles was not large enough to support a textile-machinery industry. The Mexican market had not yet reached the minimum threshold size to support capital-goods producers. When it did, later on in the century, changes in the size of scale economies in capital goods industries again put them out of Mexico's reach, a subject we will return to in the next section. Reassessing the obstacles to growth, 1880-I9io

Beginning in the i88os the obstacles which had earlier limited Mexican industrialisation began to be removed. The spark that set off this process was the inflow of capital from the United States and Europe. During the last decades of the century foreign capital and foreign entrepreneurs (who possessed knowledge about specific technologies and markets which Mexican capitalists did not) flowed into the nation, draining and retimbering the mines, spurring the growth of commercial agriculture, developing the oil industry and financing the whirlwind construction of a national railway system. By 910o, according to the available estimates, foreigners had invested close to $2 billion in Mexico's railroads, mines and a variety of other undertakings - a sum that accounted for between 67 and 73 % of the total capital invested in the country.21 The greatest part of this foreign investment was clustered in export-related enterprises. The key area into which foreign capital flowed was the transport sector. In 1873 Mexico possessed only 572 kilometres of railroads; by 1883 it had over 5,ooo. The network expanded to over io,ooo kilometres by i893 and to I6,ooo in 1903. In I910, just before the Revolution broke out, the Mexican rail system boasted over 19,000 kilometres of track. These figures

account only for track laid under federal concession. In addition, commuter and feeder lines constructed under state or municipal concessions accounted for an additional 7,8o0 kilometres.22 Although the system was laid out without any central plan, the government generally awarding concessions on an ad hoc basis, the ultimate effect was that a fairly well-integrated grid developed. The primary purpose of the railways was to move raw materials to the coast or to the northern border for export to foreign markets, but the sheer number of feeder lines built eventually gave rise to an interconnected grid that linked internal markets as well as the mining areas and the ports. By the turn of the century most of the major cities were connected to one another by rail. With the arrival of the railroad, transport costs fell precipitously, 21

Rodney Anderson, Outcastsin theirOwnLand: MexicanIndustrialWorkers,I906-1911

22

pp. 36 and 40. Coatsworth,GrowthAgainstDevelopment,

(Dekalb, Illinois, 1976), p.

19.

Industrialisationin Mexico

I

stimulating the revival of the rest of the economy. According to the conservative estimates of John Coatsworth, freight rates fell from 10 cents per ton kilometre (using wagon transport) in 1878 to 2.3 cents per ton kilometre (using railroads) in 91o0. If Mexico had tried to move the volume of freight transported by the railroads in 910ousing the next-best available technology, that is, wagon transport, the cost would have increased between five and ten times.23 To give a concrete example: the cost of shipping a ton of cotton textiles the roughly 130 miles from Mexico City to Queretaro declined from $6i in I877 to $3 in i9io.24

The fall in transport costs rebounded throughout the economy as commercial producers expanded their output to serve the now vastly wider markets they faced. Foreign investment in railroads made investments in mining, petroleum and commercial agriculture economically viable. Over the period from 1880 to 910o these economic sectors boomed: silver production increased nearly fourfold, the output of industrial metals tripled and petroleum production increased from scarcely more than 5,00o barrels in

I900

to over 8 million barrels by I9io.25

Changes in the legal environment, facilitating capitalist enterprise, occurred as well. New laws were also written to encourage foreign investment in other areas. A new mining code was written in 1887, for example, which deliberately did not reserve ownership of the nation's subsoil rights for the state, as the earlier mining code did. Instead it gave ownership to the private mining companies. In addition, in 1889 a general incorporation law was passed, allowing for ease in forming limited liability, joint-stock corporations. Other important legal changes occurred as well, including a new commercial code, banking and credit laws, public debt reorganisation and an end to internal tariffs.26 The vehicle for this process of institutional and economic transformation was Porfirio Diaz. During his 34 years in power (I876-19I

I) he

dedicated himself and the power of the state to the expansion of the economy - at whatever cost. The aim of the Diaz government was to create the internal political and economic conditions that would attract the foreign capital needed to modernise Mexico. This inflow of capital in turn provided the financial resources that permitted the continued expansion and consolidation of the central government. Under the lemma of 'Order and Progress', the Diaz regime moved to eliminate the political opposition, break the power of the regional caudillos, curtail banditry and prevent the organisation of the working class. 'Order' was to be provided 23 25

26

24 Ibid., pp. 97-IO3. Anderson, Outcasts in their Own Land, p. iz. Seminario de Historia Moderna de Mexico, Estadisticas economicasdel Porfiriato fuerga de trabajoy actividadeconomicapor sectores(Mexico, 1965), pp. 136-8 and I43. Coatsworth, 'Obstacles to Economic Growth', p. 99.

I2

Stephen H. Haber

by the federal government, often at bayonet point. The 'progress' part of the equation would be brought about by the infusion of foreign capital into the economy. This was, therefore, a reflexive process. Just as Dfaz needed to attract foreign capital in order to fuel the growth of federal government power, foreign capitalists understood that it was in their interest that a strong centralised government be created that could instil and inspire popular loyalty and maintain order and stability.27 In short, during the last decades of the century the institutional and transportation obstacles which had earlier impeded economic growth were struck down. Peasant agriculture began to give way to higherproductivity commercial agriculture; the mining sector, in decline throughout the nineteenth century, began to set new records for production and sales, and new industries, like petroleum, were founded and grew rapidly. The overall result was that the economy began to grow: per capita national income grew from $62 (I95o dollars) in 1877 to $91 in I895 and $I32

in I9I0.28

From the point of view of the Porfirian elite these economic and political changes created the conditions necessary to foment large-scale industrial investment. They controlled a state that would support their endeavours and keep the working class under control. At the same time, the creation of that expanding, wage-earning work force,29 the development of a national railroad system, the ending of provincial autonomy and internal tariffs, and the curtailment of banditry, meant that the internal market was growing rapidly. The expansion of commerce, both internal and external, was fuelling the process of capital accumulation: Mexico's merchants were growing increasingly wealthy, commanding capital that could be invested in all manner of economic activities. Most important, Mexico's capitalists believed that the process of economic growth that the country had been experiencing since the i88os was going to continue. Influenced by the neo-positivist notions of progress then popular throughout Latin America, and convinced that the Dfaz dictatorship was going to convert Mexico into a 'modern' nation, they were willing to sink their fortunes into industrial ventures. Mexico's second round of industrialisation had begun. Beginning in the i89os Mexican industry was transformed. A manufacturing sector previously characterised by small, family-owned 27

BarbaraTenenbaum, 'Planning for Mexican Industrial Development: The Liberal Nation State, Tariff Policy, and Nationalism,

28 29

i867 to I9IO',

unpublished paper,

presentedat Conferenceof the AmericanHistorical Association, December 1983. Coatsworth, 'Obstacles to Economic Growth', p. 82. The numberof workersin non-agriculturalactivitiesincreasedthreefoldbetween i 861 and 1895 to almost two million workers in the latter year. See Anderson, Outcastsin their Own Land, p. 19.

Industrialisationin Mexico

13

and family-run firms producing for local and regional markets became increasingly characterised by large, capital-intensive, vertically integrated firms producing for the national market. This transformation held true across product lines and across regions. From cement to steel, from textiles to beer, large corporations, many of them exercising monopoly control of the market (a subject we will return to in some detail shortly), began to appear, pushing aside the smaller, regional producers that had historically been the mainstay of Mexican industry. A complete description of this transformation is beyond the scope of this article, but a few examples can give the reader an idea of how Mexican industry changed once the earlier obstacles to industrialisation were removed. Perhaps in no other area was the transformation from small, regional producers to a large, national producer as pronounced as in the steel industry. In 1o00 the first integrated steel mill anywhere in Latin America was founded in Mexico, the Fundidora Monterrey steel works. The subscribed capital of the firm was I0 million pesos (about $5 million at the contemporary rate of exchange), making it the second largest manufacturing enterprise in Mexico and thirtieth largest corporation of any type. Unlike the small foundries that had preceded it, Fundidora Monterrey was a totally integrated operation, handling all phases of steel production, from the mining of the ore to the rolling of finished products. Moreover, Fundidora Monterrey's equipment was a far cry from the antiquated technology utilised by its predecessors. Iron ore was reduced to pig iron in a Massick-and-Crook-type blast furnace capable of handling I,ooo tons of ore per day, with an output of approximately 35 tons of pig iron; and three Siemens-Martin open-hearth furnaces, with a daily capacity of 35 tons each, and one Bessemer Converter refining the pig into steel ingots. The firms also owned all the necessary rolling mills, cranes, locomotives and other machinery to turn out a variety of finished steel products.30 When the mill was running at full blast, over 2,000 workers toiled round the clock in its various departments. A similar transformation occurred in the paper industry, where one new firm, the Compafifa de las Fabricas de San Rafael y Anexas, came to control the nation's entire production of newsprint and virtually controlled the domestic production of other classes of paper goods. Founded in 1890, San Rafael y Anexas ran two mills, both in the state of Mexico. Like Fundidora Monterrey, it was a totally integrated operation: 30

Jose Luis Cecefia, Mexico en la orbita imperial (Mexico, 1973), p. 87; The Mexican Yearbook,1912 (London, 1913), p. 114; Fundidora Monterrey, 'Informe Anual, 1902', p. 46; Luis Tor6n Villegas, La industria sideriurgicapesada del norte de Mexico y su abastecimientode materiasprimas (Mexico, 1963), p. 55; Oscar Realme Rodriguez, 'La industria siderurgica nacional', unpubl. tesis de licenciatura, Universidad Nacional Aut6noma de Mexico, 1946, p. 97.

14

StephenH. Haber

it owned and operated its own haciendas, where the trees were grown, ran its own mechanical wood-pulp plant, generated its own hydroelectric power, and operated its own railroad. The total number of workers in all of these operations is unknown, but in its two paper mills alone 2,000 operatives were employed. The total paid-in capital was 7 million pesos, and annual production was roughly three times that of all its small competitors combined - roughly I 2 tons of paper per day. In fact, within a decade of its founding, San Rafael eliminated almost all of its domestic competition.31 In other capital-intensive industries, like cement, beer, dynamite and explosives, and glass, a similar process occurred: large, highly capitalised firms producing for the national market replaced the numerous small shops which had preceded them.32 They also began to drive foreignproduced goods out of the market. In beer production, for example, the nation's newer, larger, firms (the Cervecerfas Cuauhtemoc and Moctezuma) not only began to challenge the numerous local beer producers, but also replaced foreign beer imports. Beer imports fell from approximately 3 million kilos in fiscal 1889-90 to just over 5oo,ooo kilos in 1910-Ii, while demand for beer was rising. The Cerveceria Cuauht6moc's production alone in 9 0 was over 2 5 times that of the total

volume of imports.33 In more labour-intensive manufacturing, especially in consumer nondurables, a similar transformation was underway. Data from the two largest of these industries, cotton textiles and cigarettes, illustrate this change best. One of the most obvious indicators of this change in the textile industry was the tremendous growth in the number and size of factories. At the end of the earlier period (1877) there had been 92 cotton mills in operation, each employing

on average 2,753 spindles, 98 looms and 128 workers. By

1895 the mills were both more numerous and larger: there were now I o mills in operation, employing an average 3,741 spindles, 112 looms and workers. By I9I0 207 they were more numerous and larger still: there were now 123 active mills, employing 5,714 spindles, 203 looms and 260

workers on average. In other words, not only were there now roughly 3'

32

33

The Mexican Yearbook, i908 (London, 1909), p. 539; The Mexican Yearbook, Io909-o (London, 1911), p. 416; The Mexican Yearbook,1912 (London, 1913), p. II, i26; Lenz and G6mez de Orozco, La industriapapelera, p. 83.

For a more complete discussion of the transformationof these industriessee Stephen The Industrialization H. Haber, Industryand Underdevelopment: of Mexico, I890-I940 (Stanford, 1989), chs. 4 and 6. Calculatedfrom data in El Economista Mexicano,24 Dec. i898, p. 249; El Economista Mexicano,27 July 1907, p. 360; Fernando Rosenzweig, Comercioexteriorde Mexico, 1877-19II:

estadzsticaseconomicasdel Porfiriato (Mexico, 1960), p. 208.

Industrialisationin Mexico

15

one-third more mills in operation than there had been three decades earlier, but they were roughly twice the size of the earlier mills.34 There had also been an increase in productivity since the early years of the industry. Physical productivity (measured in output of pieces of cloth per active loom) increased some 379 % between 1843 and 1905, an average

rate of growth of productivity of roughly 2.5 % per year. This is a low end estimate, since it does not take into account improvements in the quality of cloth or the increase in the production of fine-weave goods, both of which were substantial. Labour productivity increased as well, though this is somewhat harder to measure in the years prior to i895 because of significant changes in the allocation of the labour force between spinning and weaving departments of the mills. An analysis of post-I895 data, however, indicates a similar increase in output per worker: labour productivity

grew

by 31 % between

1895

and 1905, a 2.7%

annual

increase. 35 Whatever the exact rate of productivity growth, it is clear that Mexico's newer, highly mechanised firms were more efficient than their older competitors and certainly were able to out-compete the nation's artisans, thereby forcing both of these traditional producers from the market. Of the nation's

150 cotton

mills in

1905,

for example,

only

123 were

operating, the other 27 (almost all of them smaller, older mills) unable to compete against the newer, more efficient mills.36 The nation's artisans fared even worse: of the 4I,000 artisan cloth producers in 1895, only remained in 19I0.37 12,000 A similar transformation occurred in the cigarette industry where three giant firms, employing automated cigarette rolling machinery from France, drove the nation's artisanal producers out of the market. In 1898 the nation's output of almost 5 million kilos of cigarettes was divided up among 766 manufactories. Ten years later, in I908, though production had increased by 76% to almost 8.7 million kilos, the number of producers had fallen by over 40 % to 437. In other words, output per establishment trebled over the ten-year period, from 6,40o kilos in 1898 to 19,819 kilos in I908. By June 191I, the last date for which data exist,

there were only 341 manufactories in operation, less than half the number in existence just 12 years earlier.38 These figures probably understate the degree to which the major producers forced out the smaller manufacturers, Calculated from Secretaria de Hacienda y Credito Puiblico, Documentos, doc. 5; Ministerio de Fomento, Estadistica del Departamentode Mexico, doc. 2; Seminario de 35 Ibid. Historia, Estadisticas economicasdel Porfiriato, p. io6. 36 Calculated from Seminario de Historia, Estadisticas econdmicas del Porfiriato, p. io6. 37 Anderson, Outcasts in their Own Land, p. 47. 38 Calculated from Secretaria de Hacienda, Boletin de estadzsticafiscal, I894-I9Ir (Mexico, various years). 34

I6

StephenH. Haber

since the data on the number of firms includes cigar manufacturers (which were not mechanised and therefore not subject to the same kind of concentration because of increasing returns to scale) as well as cigarette producers. Indeed, firm level data from the three major cigarette producers suggest that they controlled 60 % of the market by the close of the Porfiriato. Data on the labour force also indicate that the big, mechanised firms were forcing the artisanal shops out of business. In I895 there were 10,397

workers employed in cigarette production. By 910o there were but 6,893, approximately half of whom were employed by the nation's three major firms.39 If we assume that cigarette output was the same in i895 as it was in 1898, the first year for which output data are available, the data indicate that output per worker increased from 473 kilos in 1895 to I,216 kilos in 19 0 - an increase in labour productivity

of

I

57 % over the

5-year period

- the apparent result of the substitution of automated rolling machinery in the big firms for thousands of workers rolling cigarettes by hand in hundreds of small shops. These figures, like those on output per factory, also tend to underestimate the degree of change in the industry. It is unlikely that output in I895 was equal to the level of 1898 - it was almost

certainly significantly lower. Thus, if anything, the degree to which the automated machinery employed by the large firms raised labour productivity was probably greater than the 157% increase estimated here.40

Family-owned enterprises remained in Mexico. In some lines of manufacturing, like leather goods, cigars and food processing, small firms continued to dominate. Even in the increasingly concentrated cotton and wool textile industries, small firms that produced for geographically isolated areas continued to hold out. Wherever there were sizeable economies of scale or speed, or where the lack of a skilled labour force created incentives to mechanise production, however, big firms were pushing the small producers out. Because no comprehensive industrial census was ever carried out during the Porfiriato, it is not possible to calculate the percentage of industrial production attributable to large-scale manufacturing. Data from the census of 929, however, shed some light on the subject. As a percentage of total national production of manufactures, the subsectors that were becoming increasingly concentrated - steel, textiles, cement, beer, dynamite, soap, paper, glass, tobacco products accounted for just over half of all manufactures produced in 1929. I 39 Anderson, Outcasts in their OuwnLand, p. 41. 40 Calculated from Secretaria de Hacienda, Boletin de estadisticafiscal, 1894-191g (Mexico,

various years).

Industrialisation in Mexico

17

estimate that these industries comprised 56.6% of the total capital invested in manufacturing and contributed 56% of total value added.41 In terms of the accomplishments of US industrialisation, this was a very modest level of industrial development. By Latin American standards, however, Mexico's industrial growth was quite impressive. By 191o Mexico had well-developed beer, cotton and wool textile, basic chemical, iron and steel, paper, cement, shoe and boot, and cigarette industries. Brazil, the most industrialised country of South America, had textile, beverage, and leather-working industries which rivalled those of Mexico in size and complexity, but did not have any of the more complex industries, such as paper, steel, cement or basic chemicals. These were not developed until the two decades following the First World War.42 Impressive as Mexico's early industrialisation was, significant obstacles continued to stand in the way of self-reinforcing growth. These obstacles, for the most part, were internal to firms, and included the inability to realise scale economies, low labour productivity, and high fixed capital costs relative to the ability of the economy to mobilise capital. Let us examine each of these, and their consequences for the rate and structure of industrial growth, in detail. The Mexican market was still relatively shallow compared to those in the advanced industrial countries. Even after a 30-year period of rapid economic growth and urbanisation in which a sizeable wage-earning class was created, nearly three-quarters of the population continued to live in small villages. A significant segment of the population continued to function outside the market economy, producing for their own subsistence through traditional peasant agriculture. Even in areas where haciendas had swallowed up Indian communities, this did not imply that all of the dislocated peasants became wage workers; sharecropping and rental arrangements still tended to predominate. In addition, many areas of the country had not yet been integrated into the national rail transport system, with large sections of the south and long stretches along the Pacific Coast remaining outside the rail grid. Most importantly, the vast majority of the population was still quite poor. The range of manufactured products they could be expected to consume was therefore restricted. No studies have been conducted to date on average wage rates and living standards during the Porfiriato, but a number of indicators point to low and unequally distributed incomes.43 First, per capita GDP was quite low: $132 (1950 41 Calculated from data in Direcci6n General de Estadistica, Primer censoindustrialde 193o (Mexico, 1934). 42 For a discussion of the development of these industries in Brazil see Wilson Suzigan, Indlstria Brasileira: Origem e Desenvolvimento(Sao Paulo, i986). 43 The minimum wage series in Seminario de Historia Moderna, Estadisticas econdmicasdel Porfiriato, pp. 147-54, are widely cited to make this point but are of dubious value for

18

StephenH. Haber

dollars) in 191o, compared

to $807 in Great Britain and $1,035 in the

United States.44 In addition, the distribution of this income was highly skewed. A redundant labour force and highly capital-intensive methods of production suggest that most of value added went to capital, not to labour. In fact, the income of the average household was so low that the consumption of cotton cloth was highly sensitive to changes in the price of corn, which along with beans and chiles made up the largest part of the diet of Mexico's working class.45 This uneven and limited market had to combine with a relatively sophisticated and expensive technological base. Mexico now had a larger, better integrated market than it did in the I840s, but there had also been significant changes in manufacturing technologies since that time, putting many industries out of Mexico's reach. In general, these advances lowered unit costs of production by taking advantage of economies of scale and economies of speed. Thus, the optimal size (the scale of production at which unit variable costs would be minimised) of firms increased substantially. This not only prevented Mexico from entering many lines of manufacturing, but also shaped the industries which did develop in several significant ways. Since the minimum-efficient scale in capital goods production had increased, and since capital goods industries now required well-developed scientific and engineering capabilities, Mexico had little choice but to import its capital equipment.46 Thus, Mexico's blast furnaces and rolling mills came from the United States, the high-speed cigarette machinery from France, the paper-making machinery from Switzerland, the textile looms, spindles, and other equipment from England, Belgium and the United States. Although this certainly sped up the process of early industrialisation,47 this imported technology, designed for mass a variety of reasons. In the first place, Mexico did not in fact have a legal minimum wage prior to the 1930s. Secondly,the authorsdid not publish the sources or methods used, so there is no way of knowing exactly what the 'minimum wage' figures represent.It is most likely the case that the reportedfigures representthe lowest wage rate that the researcherscould find for a particularyear, which is not, to put it mildly, a systematicapproachto gathering wage data.

44 Coatsworth, 'Obstacles to Economic Growth', p. 82. 45 El EconomistaMexicano, 7 May 1904, p. II4. 46

Mexico's investment in human capital was extremely low. The educational system served only the few, with the vast majorityof the populationhaving almost no formal schooling and very low levels of literacy. In I895 only 14% of the population could

socialesdel Porfiriato read and write. See Direcci6n General de Estadistica,Estadzsticas

(Mexico, I956), p. 123.

47

AlexanderGerschenkronmade this point about the role of imported technology for in Historical late developers in Eastern Europe. See his EconomicBackwardness Perspective: A Book of Essays (Cambridge, I962), ch. I.

Industrialisationin Mexico

19

production/mass consumption economies, was inappropriate for the Mexican market. This had two ramifications. The first was that there was a severe problem of excess installed capacity in many industries. In the cement industry, for example, Mexican manufacturers were able to run at only 43 % of capacity between 1906 and 9I I. In the steel industry the situation was even worse. Although there was only one major producer, Fundidora Monterrey, it still only managed

to run at 30% of capacity between

1903 and i9io.48

Even in consumer

goods, where the capital equipment was more easily divisible and scale economies tended to be smaller, there was a capacity utilisation problem. The scarcity of consistent data makes it impossible to estimate capacity utilisation rates in these industries accurately, but reports from manufacturers and analyses conducted by the Mexican financial press on cotton textile manufacture, the largest and best developed of the consumer goods industries, indicate that the industry could not run at capacity. In competing for control of the market at the turn of the century, Mexico's major textile mills began to switch to the recently developed, high velocity, automatic machinery. If the entire Mexican textile industry had then been run at capacity, the market could not have absorbed all the output. Indeed, crises of 'overproduction' occurred on a fairly regular and basis: at least twice during the latter part of the Porfiriato, in 1901-2 again in I907-8. Throughout the years of the Revolution the 1920S the problem was endemic.49

(1910-17)

and

The second ramification of the importation of capital goods was that it increased the costs of entry: Mexican industrialists had higher start-up costs than did industrialists in the advanced industrial economies. Not only did they have to pay for the foreign-produced machinery; they also had to set aside funds to cover the cost of transport, insurance in transit, and the salaries of the foreign technical personnel who set up the plant. In the cotton textile industry, for example, these added expenses pushed up the final cost of erecting a mill in 191o from $12.72

per spindle in Great

Britain to $I9.72 per spindle in Mexico, a difference of nearly 6o %.50 As a result, a large part of Mexican industry was inefficient from its beginning. Foreign-produced capital goods led to high costs of entry and low levels of capacity utilisation, with concomitantly high unit costs of production. Had Mexico been able to export manufactures this unit cost problem could have been solved. An attempt was in fact made in 1902 to 48 p. 33. Haber, Industryand Underdevelopment, 49 See El EconomistaMexicano, I8 Jan. 1902, p. 245; 2I June 1902, p. 203; 5 Sept. 1903 (English edition), p. 536; ii July 1908, p. 297. 50 Gregory Clark, 'Why Isn't the Whole World Developed? Lessons from the Cotton Mills', Journal of EconomicHistory, vol. 47 (I987), P. I46.

20

StephenH. Haber

enter the international market, but was a resounding failure. The lack of a merchant marine to ship its goods to foreign markets, the lack of institutions to provide credit to exporters, the competition of producers from the advanced industrial economies, and high tariff barriers in other Latin American countries, all worked against Mexican industrial exports.51 A second obstacle to successful industrialisation was the low of Mexican labour with in that the advanced productivity compared industrial countries, and this again pushed up production costs. The Mexican working class had its social roots in the peasantry; many workers were only recently off the farm; some moved back and forth between the factory and the field. They therefore worked with the rhythm of a peasantry, not of an industrial proletariat. For this reason Mexican industrialists did not have the same degree of control over labour as did their counterparts in the United States, England or Germany. Though they could force workers to work long hours, they could not instil in them the attitudes and values that are essential to the development of industrial discipline. Like European industrialists in the late eighteenth and early nineteenth centuries, Mexican factory owners regularly complained about the 'laziness' of their work force and their inability to force the workers to submit to routinised work.52 This resistance on the part of the working class took two forms. One was that workers openly resisted employers' attempts to increase productivity or achieve greater discipline on the shop floor. There were even occasions when entire shifts of workers abandoned their machines in order to attend a fiesta at a nearby hacienda or barrio.53 The second was that Mexican workers generally worked less intensively than did their American or European counterparts. In the weaving departments of Mexico's cotton textile mills in 910o, for example, the average worker operated 2.5 looms,

For a detailed discussion of this attempt to export manufactured goods during the pp. 39-43. Porfiriato see Haber, Industryand Underdevelopment, 52 For a typical analysis of the productivity of Mexican labour by a Mexican industrialist see Jose Robredo, Punto de vista de los industrialesde hiladosy tejidosde la Repiblica (Mexico, 1925), p. 5I. For a similar analysis by a foreign observer see G. Jenner, 'Informe de Mr G. Jenner sobre la inversi6n del capital ingles en Mexico', in Informesy documentos relativosa comerciointerior,mes de setiembrei886 (Mexico, i886). This experience with the slow transition to 'factory time' was not unique to Mexico. In Meiji Japan, for example, people returning from study or work abroad regularly commented on the laxness and lack of time discipline of Japanese workers compared to their Western counterparts. The transition to modern conceptions of time and work, however, appears to have proceeded more rapidly in Japan than in Mexico. See Thomas C. Smith, Native Sourcesof JapaneseIndustrialization,I7fo-1920 (Berkeley, I988), pp. 225-8. 53 One of the more notorious cases of this occurred in the San Lorenzo cotton mill in Orizaba Veracruz in 1923. For details on this case see Archivo General de la Naci6n, Ramo de Trabajo, Box 560, file 6, doc. iI. 51

Industrialisationin Mexico

21

whereas British operatives worked 3.8, and New England operatives worked 8.o. In the spinning department of the mills the results were similar: in Mexico 540 ring spindles per worker, compared with 625 for British workers and 902 for New England workers. When reduced to a single measure of staffing levels - loom equivalents per worker - the data indicate that Mexican textile mills employed almost twice as many workers per machine as did British mills, and over two-and-one-half times as many workers per machine as did New England mills.54 To an extent the disadvantages of lower capacity utilisation rates, higher start-up costs, and lower labour productivity would have been offset by lower labour costs in Mexico. Workers in Mexico's textile mills, for example, earned on average roughly half of what British workers did and less than one-third of what New England workers did. But it is clearly not the case that these lower labour costs would have made up for the higher capital and operating costs in the textile industry: Mexico's costs of production were still almost 20 % higher than in Great Britain.55 Moreover, in less labour-intensive industries, like beer brewing, glass making, steel and chemicals, the savings from lower wages would have been even less significant than they were in the textile industry. Mexican manufacturers therefore found it difficult to compete in the home market against foreign manufacturers without the protection and support of the government. Almost all of Mexico's major industries received some sort of tariff protection or federal subsidy. Beginning in the i88os import duties on manufactures jumped markedly, with the schedules being revised upwards in 1892, 1893, 1896 and I906.

By the end of the

Porfiriato, levels of protection were so high that one US commercial agent reported that 'the Mexican tariff on cotton goods is the highest in the world [sic], being exceeded only by those of Russia and Brazil. In some classes of cloth the duty amounts to three times the value of the goods abroad. '56 In other product lines levels of protection were extremely high as well. In the dynamite industry, to cite but one example, a series of different import duties and excise taxes produced a tariff of 80 % on imported goods.57 In addition to protective tariffs, most of the country's major manufacturing enterprises operated under some kind of federal concession, which gave them tax-exempt status for a period of between seven 4 Clark, 'Why isn't the Whole World Developed?', pp. I5 1-2. Ibid., pp. 146 and 50o. 56 William A. Graham Clark, Cotton Goods in Latin America: Part One, Mexico, Cuba, and Central America (Washington D.C., 1909), p. 38. 57 Calculated from data in The Mexican Yearbook, 1909-I910, El Economista pp. 414-15; 55

Mexicano, 4 Jan. 1902, p. 217.

StephenH. Haber

22

and

30

years. In 1893 the government declared that all new industries

capitalised

in excess of 250,000 pesos were exempt from direct federal

taxes and from customs duties on the machinery and other materials needed to erect their factories. The required minimum capitalisation was lowered to 0oo,ooo pesos in 898.58 A few enterprises were able to do even better in their negotiations with the Diaz government. It sometimes granted firms the sole right to tax-exempt status within a particular product line. In effect, in order to promote industrialisation in an extremely difficult environment, the government was setting up officially sanctioned and subsidised monopolies. Besides the development of inefficient, highly protected industry, the other major feature of this wave of Mexican industrialisation was the tendency of manufacturing to be extremely concentrated. Within any given product line a few large firms controlled the lion's share of the market. In steel, glass, soap, paper and dynamite production, single firms held monopolies or near-monopolies. Cigarette manufacturing was dominated by two horizontally integrated giants (there were actually three firms, but the largest producer owned 5 % of the stock in the nation's second largest, effectively reducing the number of competitors to two). In beer and cement three big firms carved up the market. Even in cotton textile production, which is usually characterised by near-perfect competition, two firms claimed roughly 20 % of total national production and almost all the production of fine, high-quality goods. A comparison of four-firm concentration ratios indicates that Mexico's textile industry was 6o % more concentrated than Brazil's and 280 % more concentrated than that of the United States in I9I0.59

This non-competitive industrial structure evolved for two reasons. First, as has already been discussed, technological considerations limited the number of producers who could hope to prosper in Mexico's shallow market. Secondly, the underdeveloped nature of Mexico's financial sector meant that a small group of financiers skilled at manipulating both the market and the state held an unusual degree of political and economic power. They were therefore able not only to demand (and receive) high levels of protection from the government, but were also able to employ a wide arsenal of anti-competitive weapons designed to create barriers to 58

59

Matsuo Yamada, 'The Cotton Industry in Orizaba: A Case Study of Mexican Labor and Industrialization During the Diaz Regime', unpubl. MA diss., University of Florida, I965, p. 49. For a more complete discussion of industrial concentration and the creation of barriers to entry see Haber, Industry and Underdevelopment,chs. 4 and 6. For an analysis of concentration ratios for the textile industry see Stephen H. Haber, 'Industrial Concentration and the Capital Markets: A Comparative Study of Brazil, Mexico, and i the United States, 1830-I930', Journal of EconomicHistory, vol. 5 (199).

Industrialisation in Mexico

23

entry to would-be competitors. How then, did the obstacle of a poorly developed capital market affect the development of Mexican industry? As has already been discussed, by the time that this wave of industrialisation was underway in Mexico, the cost of entry to the market in manufacturing was much higher than it had been in late eighteenth and early nineteenth centuries when Great Britain began what came to be known as the First Industrial Revolution. By the late nineteenth century the state of manufacturing technology was such that entry into the mass production of most industrial goods required capitalisations running into millions of dollars. What had evolved in Europe over a long period of time, financed through an extended process of reinvestment of profits, now had to be purchased all at once. Entry costs were raised even further by the fact that the technology and capital goods that had to be imported were expensive to ship, to insure in transit, and to set up once they arrived in Mexico. Manufacturers therefore had to raise initial capitalisations running into the millions of pesos, but the ability of the economy to mobilise capital was restricted. A government-affiliated industrial finance bank could have been used to channel capital into manufacturing, but this had already been tried from I830 to 1842 with very limited success. The idea was never

revived, even after the conditions for industrialisation became more propitious. The private banking system could likewise not be counted on to serve as a source of finance capital. Mexico's banking system, still in its infancy, was designed to do nothing more than facilitate commerce. As late as

9Io0,

the entire banking system still numbered less than 50 firms

- most of them extremely small enterprises with capitalisations in the area of only $250,000.

Almost all of these banks were legally constituted

only

to serve as sources of short-term credit, meaning that they could at most provide working capital to manufacturing companies.60 Most of the investment capital for Mexican manufacturing therefore came from the nation's most prominent merchant-financiers. They were the only group in Mexico with sufficient liquid wealth to finance the very expensive plant and equipment that had to be imported. Given the large capital requirements of the enterprises and the perceived risk in manufacturing, no one financier would commit all his resources to any single project. Instead, several financiers would combine to form a joint stock company. Since the number of major financiers was relatively small - no more than 25 - the overall effect was that a tight clique controlled

Mexico's most important manufacturing companies. 60

For a detailed discussion of banking institutions during the Porfiriato see Hilda Sanchez Martinez, 'El sistema monetario y financiero mexicano bajo una perspectiva hist6rica: el Porfiriato', in La banca,pasadoy presente: ensayosdel CIDE (Mexico, 1983).

24

Stephen H. Haber

The two most salient characteristics of this class of merchant-financierindustrialists were that few of them were Mexican by birth and fewer still knew anything about manufacturing. Of the nine men who made up the first board of directors of the Fundidora Monterrey steel mill, for example, only one had any relevant technical experience.61 These were not the tinkerers of the English Industrial Revolution or the productionorientated engineers and scientific managers of US industry. They were financiers whose principal talents lay in making deals to maintain their monopoly positions and in manipulating the economic apparatus of the state to provide them with protection from foreign and domestic competition. Their presence on the boards of Mexico's major manufacturing enterprises therefore reinforced the tendency of Mexican industry to collude rather than compete. These merchant-financier-industrialists were well positioned to shape government policies to their liking. Indeed, they were the economic backbone of the Porfirian state: they subscribed to the government's treasury bonds, they sat on the boards of the nation's most important financial institutions and they represented the government in international financial markets when it borrowed money abroad.62In fact, it was not so much a case of the state representing the interests of these financiers as it was that these financiers were the state. They controlled the emission of paper money through their ownership of the Banco Nacional de Mexico - there was no government-run central bank. They shaped national monetary and exchange-rate policy through the seats they occupied on the Comisi6n de Cambios y Monedas, and they held the strings on the flow of international loans to the Mexican government through their connections to the major banks in Madrid, Geneva, Paris and New York. This class of merchant-financier-industrialists used their political and economic power to pursue rent-seeking strategies. At the same time, poorly developed capital markets kept other groups from challenging their hold on industry. That is, rather than make innovations in new processes or techniques as entrepreneurs in some follower countries did (Germany stands out as a case in point), Mexico's entrepreneurs sought to limit competition and earn monopoly rents. Given the fact that their entrepreneurial talents and experience were in commerce and moneylending - not in production - and given the fact that demand for many of their products was highly inelastic, this strategy made sense from the point of view of maximising returns to the firm. Indeed, returns to manufacturing were low (many enterprises, including some of the 61 Fundidora Monterrey, 'Informe Anual', 1902. 62 Mexico's most important financiers operated on an international scale and had connections to the big foreign banks. See Sanchez Martinez, 'El sistema monetario'.

Industrialisationin Mexico

25

monopolies, were perennial money-losers or made very irregular profits), and would have been lower still if firms had not pursued anti-competitive strategies.63

A detailed discussion of how Mexico's financiers carried out strategies designed to limit competition is beyond the scope of this article, but in general it usually involved closing off access to some important factor of production, like technology (through the control of foreign patent rights), raw materials or government protection. Where they could, firms also tried to close off access to the distribution and marketing networks. These strategies were usually accompanied by an attempt to buy out any competitors or would-be competitors. Since there was not a well-developed capital market in Mexico, these strategies could provide market dominance over the long run. This was not the case in countries that had well-developed capital markets, like the United State or - to cite a Latin American case, Brazil - since the high prices imposed by a monopoly attracted competitors who could use the national capital market to finance the creation of a rival firm.64 This is, in fact, what happened to most of the giant consolidations formed in the United States during the great merger movement of the I89os. An analysis of paper manufacturing in the two countries elucidates the point. In both countries single firms (the San Rafael y Anexas paper company in Mexico, and the International Paper Company [IPC] in the United States) came to control the market in the 89os through strategies of mergers and consolidations with rival firms at the same time that they integrated vertically in order both to capture the economies of speed in paper manufacturing and to cut off access to raw materials to would-be competitors. Both firms also concentrated on the production of newsprint, the product line with the largest market, (though they produced other goods as well). In the United States, the monopoly rents that IPC earned attracted competitors who were financed by the stock and bond markets. Within seven years of IPC's founding, 20 new producers entered the market, driving IPC's share of the market down from 64 % in 1900 to 48 % by 1905.65 In Mexico, the lack of a well-developed

capital market and

the cohesiveness of the merchant-financier families who dominated the 63

For an analysis of the profitability of Mexican manufacturing during this period see ch. 7. Haber, Industryand Underdevelopment, 64 Brazil had a much better developed banking system and far larger and more complex stock and bond markets than Mexico did. Over the long run this allowed Brazil to overtake Mexico's early lead in industrialisation and also gave rise to lower levels of industrial concentration than existed in Mexico. See Haber, 'Industrial Concentration'. 65 Naomi Lamoreaux, The Great Merger Movement in American Business, 1894-1-90 (Cambridge, Eng., I985), pp. i26-7.

z6

StephenH. Haber

economy enabled San Rafael to hold on to its control of the market over the long run. It was not until 1936 that San Rafael's monopoly was broken, when the Mexican government, for political reasons, decreed that the distribution of newsprint was a strategic industry that should be controlled by the State, not by a private firm. It therefore created a government distribution monopoly, though San Rafael continued to control production. In short, during the

20

years between i890 and

i910

Mexico underwent

a rapid process of economic modernisation and growth. A national market was created by the building of a rail network, the mining sector was revitalised, a petroleum industry was created, agriculture became increasingly commercialised and industry moved into a wide variety of new products that had never before been produced in Mexico, driving imported goods from the market and pushing the nation's artisans to the wall. But the contradictions and obstacles inherent in the rapid industrialisation of an economy like Mexico's prevented self-sustaining industrialisation from taking hold. The market was still too small to support capital goods industries (it could barely sustain consumer-goods producers in many lines of production), the industries that did exist were unable to compete without extensive support and protection from the government and a group of rent-seeking financiers came to control the manufacturing sector. The end of the Pax Porfiriana, I9io-j93o

It could be argued that the process of economic growth from i880 to 910o would have eventually overcome the obstacles which stood in the way of self-reinforcing industrialisation. The continued growth of the economy might have broadened and deepened the market, expanded the banking sector and formed a national capital market, built an educational system adequate to the task of developing engineering and industrial design capabilities, and continued to break down pre-modern forms of agricultural production. While this could have happened in theory, the reality was that the uneven process of economic growth in Mexico set off a revolution against the Dfaz dictatorship from both above and below, not only ending the model of growth of the Porfiriato, but also producing a slowdown in industrial growth for the next 25 years. Contrary to much of the popular mythology about the effects of the Mexican Revolution (much of it motivated, as John Womack has pointed out, more by ideological presuppositions than by an analysis of the empirical data),66 the Revolution 66

of 1910-17

did not destroy Mexico's

John Womack, 'The Mexican Economy During the Revolution, 1910-1920: Historiography and Analysis', Marxist Perspectives,vol. I (1978), pp. 80-I23.

Industrialisationin Mexico

27

industrial plant, nor did it send the industrial barons of the Porfiriatointo exile, never to return. Mexico's physical plant, especially the large firms that had the political and financial resources to protect themselves, came through the fighting in one piece. Thus, output quickly reached its Porfirian levels once the fighting stopped in 1917. Mexico's industrialists were also still present after the Revolution, and in fact were even mobilising politically to defend their interests as early as 19I7 against the new state taking shape.67 The Revolution did, however, significantly affect the confidence of Mexico's industrialists. It was not that the post-revolutionary governments were anti-industrial or anti-business - they were in fact made up of businessmen - but rather that the rules of the game had been rewritten by the Revolution. The industrialists no longer controlled the reins of power the way they had done during the Porfiriato; urban labour, which had played an important role in the defeat of the peasantry during the civil war now played a much larger role within the state, extracting

of I914-17,

important concessions in the 1917 Constitution such as the right to organise and strike, an eight-hour day, a six-day work week, and equal pay for equal work. The political climate was also unstable: it was not entirely clear what direction Mexico would follow during the I920s, whether there would be a renewed round of violence moving the Revolution to the left in an attack on private property or to the right against the progressive compromises

of the Constitution

of I917.

The result was a marked deceleration in new capital spending by industrialists, a dramatic fall in the value that financiers placed on their assets, and a jump in financial rates of return as investors began to demand sizeable risk premiums to keep their money in Mexico. An analysis of capital spending by Mexico's major steel, cotton textile and cigarette manufacturers, for example, reveals that firms began to run their already established plants into the ground after the Revolution. The book value of the physical plant of the country's major textile manufacturer (the Compafi?a Industrial de Orizaba) fell by i6 % between 1920 and 1924, and of the major cigarette producer (El Buen Tono) by 21 % between i918

and 1924. Even in the steel industry, where there was considerable new investment

between

1919 and

I921,

the brakes were soon put on new

capital spending, resulting in a drop in book value of five per cent between 192I

and

1924.

Data from the federal tax register of the cotton

textile

industry, which covers all cotton spinning and weaving firms, indicate a 67

PrimerCongreso Nacional de Industriales,Resenaj memoriadelprimercongreso nacional de industriales reunidoen la Ciudadde Mexicobajoel patrociniode la Secretariade Industria Comercioy Trabajo (Mexico, 1918).

28

StephenH. Haber

similar decline, with the value of capital in equipment and buildings per active mill falling by

3 % between

1922 and 1925.68

The lack of capital spending was driven by the pessimism of investors, who were revaluing their assets downwards. An analysis of the ratio of market-to-book values for the three firms discussed above indicates that investors

perceived that their assets were worth only 5o % of what they

had perceived them to be prior to the Revolution. Along with a decline in the value of assets came concomitant jumps in stock yields, as investors began to demand a sizeable risk premium in order to stay in the Mexican market. The average yield on common stock in Mexico's major industrial companies increased from 4.6% per annum during the period between 1896 and 1910 to 9.4 % per annum between 19 8 and

92

5. Similarly, real

financial returns to investors increased sixfold; jumping from 3% per to 8.7 % during the period from 9 8 annum during the period 9go 01-0o a to 1925, the result of dramatic fall in the real market values of stocks between

1910 and 1918, which produced

sizeable capital gains for those

willing to invest in the market afterwards.69There is also evidence that the 1920S were a period of capital flight in Mexico, as investors began to keep more of their cash assets abroad.70 The pessimism of the investment community in Mexico was borne out in the latter part of the 1920S as Mexico began to enter the Great Depression. Beginning in 1926 the prices of Mexican exports went on a downhill

roller coaster ride. Between

1926 and I928 export revenues fell

by ten per cent, from $334 million to $299 million. The drop accelerated thereafter, with exports eventually falling to only $97 million in 1932. This drop was not compensated for by a similar drop in import prices. Not only did the volume of exports decline by 37 % between 1929 and 1932, but the

terms of trade deteriorated by 21 % as well. Thus, the purchasing power of Mexican exports fell by over 5o% in just three years. The negative effects of the collapse of the export sector, which had significant effects on employment and demand, were exacerbated by pro-cyclical monetary and fiscal policies.71 Though the estimates are rough, the available data indicate that real per capita Gross Domestic Product fell by 5.9 % in 1927, 0.9% in 1928, 5.4% in 1929, and 7.7% in 1930. In I93 there was a slight recovery, with an increase in GDP of 1.5 %, but it was not long-lasting. 68 For a detailed discussed of the effects of the Revolution, as well as a discussion of how ch. 8. these values were calculated, see Haber, Industr, and Underdevelopment, 69 For a detailed discussion of the sources and methods employed in this analysis see Haber, Industryand Underdevelopment, pp. 144-8. 70 Joseph Sterrett and Joseph Davis, The Fiscal and Economic Conditionof Mexico (New York, i928), p. i90. 71 Enrique Cardenas, La industrialiiacionmexicanadurantela Gran Depresidn(Mexico, 1987), chs. 3 and 4.

29

Industrialisationin Mexico

In 1932 real per capita GDP fell by an incredible I6 %. Over the six-year period real GDP per capita fell by 30.9%.72 The results, as one might

expect, of this downturn were contractions in industrial output, nominal wages, employment, profits and new investment. In fact, many of Mexico's major industrial firms came close to bankruptcy during the depression.73 It was not until the mid-I93os, therefore, once a general economic recovery began, that a new round of industrial development took place. Several factors were behind this recovery. One was government deficit spending, which not only pushed up aggregate demand, but also served to remove some of the remaining obstacles to industrialisation. Extremely significant here was spending on a federal highway network (the road system increased from 1,426 kilometres in 1930 to 9,929 kilometres by i940),74 which created demand for producer goods like cement and steel,

as well as linked markets, and spending on irrigation systems in the north of the country, which increased agricultural productivity. The second was the agrarian reform, which redistributed wealth in the countryside and therefore, over the medium term, redistributed incomes as well, thereby increasing the size of the market. Thirdly, though there was a good deal of conflict between

the government

of Lazaro Cardenas (1934-40)

and

industrialists, one of Cardenas' accomplishments was the restoration of social peace in Mexico. Industrialists may not have wielded political power in the Cardenas government, but Cardenas did create a stable political environment through the formation of a national party, thereby restoring the confidence of the industrialists in the viability of the State. Mexican manufacturers no longer had to fear the vagaries of an unstable political system. Finally, there was a recovery of the price of Mexico's major exports, silver and oil, which revived the economy through the same mechanisms that had earlier forced it to contract. Overall, from 1933 to 1939 real per capita GDP grew by year.7

24

%, or roughly three per cent per

With the increase in demand, the return of profits, and the creation of a stable political system, industrialists, Porfirian as well as a new group of immigrant entrepreneurs, began a new round of investment in manufacturing. Not only did investors revalue their assets upwards (the ratio of market-to-book values increasing twofold for the older, major firms between

1932 and 1938), but spending

on new plant and equipment

72 Calculated from data in Instituto Nacional de Estadistica Geografia e Informatica, Estadisticas historicasde Mexico (Mexico, I985), P. 311. 73 For a detailed discussion of the effects of the depression on employment, wages, profits, ch. 9. and investment see Haber, Industryand Underdevelopment, 74 Instituto Nacional de Estadistica Geografia e Informatica, Estadisticas historicas de 75 Calculated from ibid., p. 311. Mexico, p. 566. LAS 24

Stephen H. Haber

30

occurred as well. On an aggregate level, real expenditures on private fixed capital formation were I67 % higher in

9 39 than they had been in

1932.

Part of this increase in capital spending was the result of modernisation programmes undertaken in older industries. The value of the physical plant of the Fundidora Monterrey steel mill, for example, increased by 35 % between 1934 and 1937. The value of the physical plant of the major

cotton textile producer, the Compafifa Industrial de Orizaba, almost doubled during the same period, increasing by 98 %.76 Of equal significance to the renewed investment programme of older industries was the creation of new firms. Basically, these new companies did not enter into competition with the old established giants. They left beer brewing, steel-making, paper milling, and other capital-intensive, vertically integrated operations to the companies that had dominated those lines of manufacture since the Porfiriato. Instead, the new entrants to the market concentrated on new products, like rayon knitwear, cotton knitwear, silk and high-grade cotton shirtings, hosiery and other goods that had been imported up until then. In fact, of the 692 firms listed in the 1938 tax registry of the textile industry, well over half were involved in the manufacture of these new products.77 Manufacturing, in fact, came to lead the economy at this point, with value added in this sector growing I45 % faster than GDP as a whole.78 In fact, for the next forty years manufacturing was the engine of growth of the Mexican economy. In this round of Mexican industrial development new obstacles to growth presented themselves, and many of the old obstacles continued to persist. As had occurred in the earlier period, these shaped both the rate and structure of industrialisation, and prevented, once again, a self-sustaining process of industrialisation from taking hold. Conclusions

Although Mexico did not become an industrial society until the second half of the twentieth century, a significant amount of industrialisation was achieved in the nineteenth century - particularly during the 20 years from I890 to

1910.

Even as early as the I83os Mexican

policy-makers

and

entrepreneurs attempted to copy the experience of Western Europe by introducing modern technology and modern organisational arrangements in the cotton-textile industry. Mexico's low level of nineteenth-century industrial growth was not therefore a product of not wanting to industrialise. 76 Cardenas, La industrialiacio'nmexicana, p. 144; Haber, Industryand Underdevelopment, ch. 10.

77 Secretaria de Hacienda y Credito Publico, 'Directorio de las fibricas de hilados y tejidos registrados' (Mexico, I938). 78 Cardenas, La industrialiacion mexicana, p. 1o.

3

Industrialisationin Mexico

The problem for Mexican industry during the early and mid-nineteenth century was that the economic environment was not conducive to manufacturing growth on a large scale. Internal tariffs, the lack of reliable and inexpensive transport and widespread brigandage precluded the creation of a national market. Insecure property rights, a corrupt judiciary and political instability made contracts difficult to enforce and gave rise to an uncertain institutional environment. Finally, low household incomes precluded the development of deep and secure markets. During the three decades associated with the dictatorship of Porfirio Diaz these obstacles were overcome, but by this point new constraints on industrial growth presented themselves. These obstacles were largely internal to firms. In the nearly ioo years since the onset of the Industrial Revolution important developments had occurred in manufacturing technology, which both raised the minimum-efficient scale of production and increased fixed capital costs. The Mexican market, in the meantime, had not grown in step with this technological revolution, nor had there been the kind of change in labour productivity and attitudes about work that had occurred in Western Europe. In addition, Mexico had not created institutions designed to mobilise the capital needed to finance this new technology. Mexico was therefore unable to adapt much of this new technology without either creating non-competitive production and marketing arrangements or resorting to government subsidies and protection. The result was an impressive amount of industrialisation by Latin American standards, but a relatively unimpressive industrial revolution by the standards of the North Atlantic economies. It is conceivable that had the model of growth of the Porfiriatopersisted, many of these obstacles to industrialisation might have been overcome. Incomes might have risen, educational levels might have improved, the banking system would have continued to expand and the internal market would have grown. Continued economic growth might have created its own demand, eliminating the obstacles that stood in the way of industrialisation. Indeed, Mexico's industrial entrepreneurs were banking on this to happen. It was for this reason that they often invested ahead of the market, erecting industries which were too large for the demand they faced and losing money over the short run in the hope that when the market expanded they would be in a position to take advantage of it. This, however, did not happen. In fact, the very inequalities that were an integral part of the Porfirian model of economic growth produced a Revolution. The Revolution, while it did not destroy the nation's industrial plant, created significant uncertainty among the industrialist class. These events, coupled with the onset of the Great Depression, produced a slowdown in the rate of growth of industrial investment until 2-2

32

StephenH. Haber

the middle of the 1930s. It was therefore not until the middle of the twentieth century, almost a century after the first rumblings of industrialisation began in Mexico and nearly 50 years since the first largescale industrial enterprises were established, that Mexico began to make the transition to an industrial society in which manufacturing led the economy.