Hostname: page-component-78c5997874-fbnjt Total loading time: 0 Render date: 2024-11-05T08:06:14.455Z Has data issue: false hasContentIssue false
  • English
  • Français

Political Instability And Economic Performance: Evidence from Revolutionary Mexico

Published online by Cambridge University Press:  13 June 2011

Stephen Haber  and
Armando Razo
Stephen Haber
Affiliation:
Stanford University
Armando Razo
Affiliation:
Stanford University
Get access

Abstract

What is the relationship between political instability and economic growth? This article examines this question using insights from the new institutional economics applied to a canonical case: revolutionary Mexico. The authors argue that investment and growth during and after the revolution were less affected by political instability than one might predict based on the extant theoretical and empirical literature. They find that while investor expectations, output, and productivity were sensitive to the interdiction of factor and product markets during the years of intense revolutionary violence (1914–17), the periods of political instability before (1910–13) and afterward (1918–34) had relatively minor effects on investor confidence, investments in new plant and equipment, rates of entry and exit by firms, industrial structure, and productivity growth.

Type
Research Article
Information
World Politics , Volume 51 , Issue 1 , October 1998 , pp. 99 - 143
Copyright
Copyright © Trustees of Princeton University 1998

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Following Alexander George, we employ revolutionary Mexico as a “disciplined-configurative” case study, whose purpose is to evaluate and build general theories. For a discussion of the use of single historical cases in theory building, see George, Alexander L., “Case Studies and Theory Development: The Method of Structured, Focused Comparison,” in Lauren, Paul Gordon, ed., Diplomacy:New Approaches in History, Theory, and Policy (New York: Free Press, 1979)Google Scholar. In general, there is no agreement on appropriate definitions of revolutions and political instability. Scholars such as Skocpol have argued that there are fundamental differences between political and social revolutions. See Skocpol, Theda, States and SocialRevolutions:A ComparativeAnalysis ofFrance, Russia, and China (Cambridge: Cambridge University Press, 1979), 4–5CrossRefGoogle Scholar. This paper provides an analytical framework with consistent definitions of these terms for systematic studies of the relationship between instability and growth. Our point of departure is the idea that the transformation of institutions in most societies is endoge-nized within the existing political system. When conflict can no longer be mediated peacefully and without regime change, we have conditions, to be made explicit below, that may lead to instability and revolutions.

2 Our data sets include information from virtually all of the extant mechanized manufacturing industries in Mexico during the period under study, including the steel, cement, tobacco products, paper, beer brewing, cotton and wool textile, and soap and glycerin industries. Because the data sets pertaining to Mexico's largest and most geographically dispersed manufacturing industry—cotton textiles—are particularly detailed, we focus much of our discussion on the econometric analysis of that industry.

3 See, for example, Dunne, John, Modern Revolutions:An Introduction to theAnalysis ofa PoliticalPhe-nomenon (Cambridge: Cambridge University Press, 1972), 3Google Scholar; Skocpol (fn. 1), 175–205, 218–29, 265–74; Moore, Barrington Jr., Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making ofthe Modern World (Boston: Beacon, 1967), 40–110Google Scholar.

4 These studies also tend to select cases on the dependent variable, limiting their ability to explain the causes of revolutions. See Barbara Geddes, “How the Cases You Choose Affect the Answers You Get: Selection Bias in Comparative Politics,” in James A. Stimson, ed., PoliticalAnalysis: An Annual Publication of the Methodology Section of the American Political Science Association, vol. 2 (Ann Arbor: University of Michigan Press, 1991).

5 Samuel Amaral and Leandro Prados de la Escosura, eds., La Independencia LatinoAmericana: Con-secuencias Econdmicas (The economic consequences of Latin American independence) (Madrid: Alianza Editorial, 1993).

6 See, for example, Londregan, John B. and Poole, Keith T., “Poverty, the Coup Trap, and the Seizure of Executive Power,” World Politics 42 (January 1990)CrossRefGoogle Scholar; Alesina, Alberto et al. , “Political Stability and Economic Growth,” Journalof Economic Growth 1 (June 1996)Google Scholar; Robert J. Barro, “Economic Growth in a Cross Section of Countries,” Quarterly Journal of Economics (May 1991); Londregan, John B. and Poole, Keith T., “The Seizure of Executive Power and Economic Growth: Some Additional Evidence,” in Cukierman, Alex, Hercowitz, Zvi, and Leiderman, Leonardo, eds., Political Economy, Growth, and Business Cycles (Cambridge: MIT Press, 1992)Google Scholar.

7 Levine, Ross and Renelt, David, “A Sensitivity Analysis of Cross-Country Growth Regressions,” American Economic Review 82 (September 1992), 943Google Scholar.

8 Alesina et al. (fn. 6), 191–92.

9 Ibid., 190; Barro (fn. 6), 432. Londregan and Poole note that it may not be political instability that gives rise to slow rates of growth, as much as it is the presence of unconstitutional rulers. But the causality issue is still not resolved, as it is generally the case that unconstitutional rulers often come to power in times of political instability. See Londregan and Poole (fn. 6, 1992). Some political scientists, while noting an unclear link between instability and growth, have argued in the opposite direction that growth produces instability through factors that increase social frustration at a faster rate than the rate of the increase in material well-being. See Huntington, Samuel P., Political Order in Changing Societies (New Haven and London: Yale University Press, 1968), 49–59Google Scholar.

10 Levine and Renelt (fn. 7), 943.

11 Barro, for example, employs the number of revolutions and coups per year and the number of political assassinations per million population as measures of instability, and then goes on to “interpret [these] variables as adverse influences on property rights.” See Barro (fn. 6), 432. Other investigators have tried to refine these measures. See Alesina et al. (fn. 6).

12 North, Douglass C., Institutions, Institutional Change, and Economic Performance (Cambridge: Cambridge University Press, 1990), chap. 1CrossRefGoogle Scholar; North, Douglass C. and Weingast, Barry R., “Constitutions and Commitment: The Evolution of Institutions Governing Public Choice in Seventeenth Century England,” Journal of Economic History 44 (December 1989), 808Google Scholar. Hurwicz, Leonid, “Toward a Framework for Analyzing Institutions and Institutional Change,” in Bowles, Samuel, Gintis, Herbert, and Gustafsson, Bo, eds., Markets and Democracy: Participation, Accountability, and Efficiency (Cambridge: Cambridge University Press, 1993)Google Scholar; Greif, Avner, “Microtheory and Recent Developments in the Study of Economic Institutions through Economic History,” in Kreps, David M. and Walk's, Kenneth F., eds., Advances in Economic Theory, vol. 2 (Cambridge: Cambridge University Press, 1997)Google Scholar.

13 See, for example, Alston, Lee J., Eggertsson, Thrainn, and North, Douglass C., eds., Empirical Studies in Institutional Change (Cambridge: Cambridge University Press, 1996)CrossRefGoogle Scholar.

14 Rosenthal, Jean Laurent, The Fruits of Revolution: Property Rights, Litigation, and French Agriculture, 1700–1860 (Cambridge: Cambridge University Press, 1992)CrossRefGoogle Scholar; North and Weingast (fn. 12). For a critique of North and Weingast, see Clark, Gregory, “The Political Foundations of Modern Economic Growth: England, 1540–1800,” Journal ofInterdisciplinary History 26 (Spring 1996)Google Scholar.

15 Informal institutions, such as social norms and customs, although created outside of the political system, can sometimes be as important as formal institutions if the formal political system alone does not preclude them.

16 This is only a simplifying assumption that facilitates the analysis. One can incorporate international aspects by including foreign elements in the set of agents operating under a particular system. We can also include other international factors through exogenous shocks, described below under political instability, such as the support of a foreign government to a particular faction fighting for a regime change.

17 Our framework is not exclusive of rules—for example, constitutional restraints on the executive branch—that regulate a government after its selection. These could be included under the second group of rules where the constituents may agree to be governed by a government if the latter conforms to an agreed-upon set of rules (a constitution). That is, the governments—if we assume it has its own objectives—can be one of the agents whose actions are being mediated through the political system, meaning, that the political system provides rules for the interaction (and permissible behavior) of government and society. Our distinction between a political system and government is similar to the conventional distinction between regimes and governments. However, we prefer the term political system because it reflects the ongoing interaction of its elements as opposed to the static definition of regime as a form of government.

18 Self-enforcement implies only that there are no incentives to deviate, not that everyone is satisfied with the existing political system. It can occur, for example, under two extreme sets of political conditions: (1) all agents, free to make their own decisions, find it advantageous to sustain the system (that is, cooperation is a Nash equilibrium in a game where the different players agree on the system that governs them); or (2) the government, possessing an effective enforcement technology, may impose severe sanctions on players who deviate from the rules (as in a dictatorship). Notice that this is not a game that defines static rules but rather one where agents agree on how to mediate their conflicting views in the present and future periods. Since this is essentially a repeated game, we also employ the term “stable” in a dynamic sense, meaning that the political system is in a steady state of “cooperation” such that small deviations (brought about by internal or external shocks to the system) do not affect its long-run equilibrium. Thus, we will use the terms “stable political system” and “political equilibrium” interchangeably.

19 This is not to say that endogenous institutional change does not have an effect on the economy. Even if we assume rational expectations, institutional changes—economic policy, in particular—may have a real effect on the economy if, for example, the government has an informational advantage over private agents or if there are nominal price rigidities. Our statement is simply one of forward-looking agents that are constantly adapting to their perceived institutional environment.

20 Notice that our definitions of revolution focus on procedures, which is consistent with our analysis of institutions as rules that constrain social behavior. Thus, exogenous institutional change is revolutionary only because it is not mediated through the existing system. This is a definition that can be used more systematically than one that focuses on the extent and direction of change, in which there is substantial disagreement about the threshold levels beyond which changes can be classified as revolutionary. Most importantly, our definition links revolutions and instability in a consistent and systematic way. By contrast, other definitions of revolutions as drastic institutional changes preclude the joint study of stability issues simply because any stable political system is capable in principle of producing such changes.

21 The first two channels can be thought of as direct effects since they stem directly from the existence of instability. The second two are indirect effects since they may occur under other circumstances as well.

22 North (fn. 12), 33–34; Clark (fn. 14), 564.

23 It is in this sense that we use the terms “political uncertainty” and “political instability” interchangeably: if agents are unable to assess the likelihood of many possible future governments and/or institutional arrangements (as in Knightian uncertainty), the stability of the political system can no longer be preserved.

24 Tullock, Gordon, “The Welfare Costs of Tariffs, Monopolies, and Theft,” Western Economic Journal 5 (June 1967)Google Scholar; Krueger, Anne O., “The Political Economy of the Rent-Seeking Society,” American Economic Review 64 (June 1974)Google Scholar; Posner, Richard A., “The Social Costs of Monopoly and Regulation,” Journal of Political Economy 83 (August 1975)CrossRefGoogle Scholar; Krueger, Anne O., “Virtuous and Vicious Circles in Economic Development,” American Economic Review 83 (May 1993)Google Scholar; Shleifer, Andrei and Vishny, Robert W., “Corruption,” QuarterlyJournal ofEconomics 108 (August 1993)Google Scholar; Murphy, Kevin M., Shleifer, Andrei, and Vishny, Robert W., “Why Is Rent-Seeking So Costly to Growth?” American Economic Review 83 (May 1993)Google Scholar; Mauro, Paolo, “Corruption and Growth,” Quarterly Journal of Economics 10 (August 1995)Google Scholar.

25 Clark (fn. 14), 565.

26 Perry, Laurens Ballard, Juarez and Diaz: Machine Politics in Mexico (Dekalb: Northern Illinois University Press, 1978)Google Scholar; Katz, Friedrich, “Mexico: Restored Republic and Porfiriato, 1867–1910,” in Bethell, Leslie, ed., The Cambridge History ofLatin America, vol. 5, 1870–1930 (Cambridge: Cambridge University Press, 1986)Google Scholar.

27 Coatsworth, John H., “Obstacles to Economic Growth in Nineteenth Century Mexico,” American Historical Review 83, no. 1 (1978)CrossRefGoogle Scholar; Salvucci, Richard J., “Mexican National Income in the Era of Independence, 1800—45,” in Haber, Stephen, ed., How Latin America Fell Behind: Essays on the Economic Histories of Brazil and Mexico, 1800–1914 (Stanford, Calif.: Stanford University Press, 1997)Google Scholar. Government estimates of GDP can be found in Instituto Nacional de Estadi'stica, Geografia, e Informatica (INEGI), Estadisticas Historicas de Mexico (hereafter cited as Estadisticas Histdricas) (1994), 401–2.

28 For a discussion of these institutional reforms, see Edward Beatty, “The Political Basis of Industrialization in Mexico before 1911” (Ph.D. diss., Stanford University, 1996); Holden, Robert H., Mexico and the Survey of Public Lands: The Management ofModernization, 1876—1911 (DeKalb: Northern Illinois University Press, 1994)Google Scholar; Noel Maurer, “Finance and Oligarchy: Banks, Politics, and Economic Growth in Mexico, 1876–1928” (Ph.D. diss., Stanford University, 1997); Coatsworth (fn. 27); Enrique Cardenas, “A Macroeconomic Interpretation of Nineteenth-Century Mexico,” in Haber (fn. 27).

29 For discussions of the Porfirian economic transformation, see Haber, Stephen H., Industry and Underdevelopment: The Industrialization ofMexico, 1890—1940 (Stanford, Calif.: Stanford University Press, 1989)Google Scholar; idem, “Assessing the Obstacles to Industrialization: The Mexican Economy, 1830–1940,” Journal of Latin American Studies 24, no. 1 (1992)Google Scholar; Coatsworth, John H., Growth against Development: The Economic Impact of Railroads in Porfirian Mexico (DeKalb: Northern Illinois University Press, 1981)Google Scholar; Maurer (fn. 28); Beatty (fn. 28); Ficker, Sandra Kuntz and Riguzzi, Paolo, eds., Ferrocarriles y vida economica en Mexico, 1850—1950 (Railroads and economic life in Mexico, 1850–1950: From late boom to early bust) (Toluca, Mexico: El Colegio Mexiquense, 1996)Google Scholar; Ficker, Sandra Kuntz, Empresa extranjeray mercado interno: el Ferrocarril Central Mexicano, 1880—1907 (Foreign enterprise and internal markets: The Mexican central railroad, 1880–1907) (Mexico City: El Colegio de Mexico, 1995)Google Scholar; Cardenas (fn. 28); Bernstein, Marvin D., The Mexican Mining Industry, 1850–1950:A Study ofthe Interaction of Politics, Economics, and Technology (Albany: State University of New York, 1964), 17–94Google Scholar; Reynolds, Clark W., The Mexican Economy: Twentieth Century Structure and Growth (New Haven: Yale University Press, 1970), 15–26Google Scholar.

30 See Knight, Alan, The Mexican Revolution (Cambridge: Cambridge University Press, 1987)Google Scholar; Gilly, Adolfo, La revolucidn interrumpida: edicidn corregiday avmentada (The interrupted revolution: Revised and expanded edition) (Mexico City: Ediciones Era, 1994)Google Scholar; Hart, John Mason, Revolutionary Mexico: The Coming and Process of the Mexican Revolution (Berkeley: University of California Press, 1987)Google Scholar; Womack, John Jr., Zapata and the Mexican Revolution (New York: Alfred A. Knopf, 1969)Google Scholar; Katz, Friedrich, The Secret War in Mexico: Europe, the United States and the Mexican Revolution (Chicago: University of Chicago Press, 1981)Google Scholar; John Womack, Jr., “The Mexican Revolution, 1910–1920,” in Bethell (fn. 26).

31 Ojeda, Leticia Gamboa, “La CROM en Puebla y el movimiento textil en los Anos 20” (The CROM in Puebla and the textile workers movement in the 1920s), in Memorias del Segundo Encuentro sobre his-toria del movimiento obrero (Proceedings of the Second Meeting of the History of the Workers Movement) (Puebla, Mexico: Universidad Autonoma de Puebla, 1984), 41Google Scholar; Benjamin, Thomas and Wasserman, Mark, eds., Provinces of the Revolution: Essays on Regional Mexican History, 1910—1929 (Albuquerque: University of New Mexico Press, 1990)Google Scholar; Wasserman, Mark, Persistent Oligarchs: Elites and Politics in Chihuahua, Mexico, 1910–1940 (Durham, N.C.: Duke University Press, 1993)Google Scholar; Falcon, Romana, El agrarismo en Veracruz: La etapa radical (The agrarian movement in Veracruz: The radical years) (Mexico City: El Colegio de Mexico, 1977)Google Scholar; idem, Revolucidny caciquismo: San Luis Potosi, 1910–1938 (Revolution and the political bosses: San Luis Potosi) (Mexico City: El Colegio de Mexico, 1984)Google Scholar.

32 Rancaño, Mario Ramirez, Burguesia textily politico en la Revolución Mexicana (The textile bourgeoisie and the politics of the Mexican Revolution) (Mexico City: Universidad Nacional Autónoma de México, 1987), 196–206Google Scholar.

33 The absence of effective enforcement of formal institutions dates back to earlier periods. For an early account of enforcement of property rights, see Armando Razo, “Land Policy in Colonial Mexico: Exploring the Links between Institutional Development and Economic Performance” (Manuscript, Stanford University, 1996).

34 Bortz, Jeffrey, “‘Without Any More Law Than Their Own Caprice’: Cotton Textile Workers and the Challenge to Factory Authority during the Mexican Revolution,” International Review ofSocial History 42, pt. 2 (August 1997), 256–57Google Scholar.

35 Maurer (fn. 28), 240 and chap. 6. In fact, the owners of the banks and the owners of the factories were often one and the same group. For an account of their connections, see Rancano (fn. 32), 95–130.

36 Rancaño (fn. 32), 209–17.

37 Anderson, Rodney D., Outcasts in Their Own Land: Mexican Industrial Workers, 1906–1911 (Dekalb: Northern Illinois University Press, 1976), 36Google Scholar.

38 For instance, Luis Morones, the head of the CROM (the national labor federation) was, in fact, the minister of industry, commerce, and labor, a position he used to enrich himself by extracting rents from both factory owners and factory workers. State governors, recognizing the importance of their labor constituencies, also became champions of the labor movement. Rancafio (fn. 32), 201. Bortz, Jeffrey, “The Genesis of the Mexican Labor Relations System: Federal Labor Policy and the Textile Industry, 1925–1940,” The Americas 52 (July 1995)CrossRefGoogle Scholar; Bortz (fn. 34), 253–88.

39 For a summary and critique of this view, see Womack, John Jr., “The Mexican Economy during the Revolution, 1910–1922: Historiography and Analysis,” Marxist Perspectives 1, no. 4 (1978)Google Scholar.

40 More evidence in support of our hypothesis that agents can insulate themselves from the effects of instability comes from other lines of activity not examined in this paper. For example, petroleum extraction or mining and smelting were capital-intensive, foreign-owned enclaves that had the ability to hire mercenary armies to protect their property rights and to bring to bear diplomatic and economic pressure from foreign governments, thereby mitigating the effects of political instability. In fact, Mexican crude petroleum output grew more than fortyfold during the revolution. Estadisticas Historicas (fn. 27), 559. For the history of foreign petroleum companies in Mexico, see Brown, Jonathan C., Oil and Revolution in Mexico (Berkeley: University of California Press, 1993)Google Scholar; and Hall, Linda B., Oil, Banks, and Politics: The United States and Postrevolutionary Mexico, 1917—1924 (Austin: University of Texas Press, 1995)Google Scholar.

41 Haber (fn. 29, 1989), chap. 4; Veronica de Allende Costa and Luis Felipe López Calva, “La economfa mexicana durante el Porfiriato: análisis macroeconomico e interactión entre los sectores publico y privado (The Mexican economy during the Porfiriato: Macroeconomic analysis and the interaction of the public and private sectors) (B.A. thesis, Universidad de las Americas-Puebla, 1991), 198–203.

42 Haber (fn. 29,1989), 91–92, 143, 188.

43 Data sources on the cotton textile industry are as follows: 1837–39 and 1841–42 data were obtained from Secretaría de Hacienda y Crédito Público (SHCP), Documentospara el Estudio de la Industrializatión de México, 1837–1845 (Mexico City: SHCP, 1977), 82; data for 1840 were partially retrieved from L. Barjau Martinez et al., “Estadísticas Económicas del Siglo XIX,” Cuadernos de Trabajo del Departamento de Investigations Históricas, Institute National de Antropologie é Historia (INAH) no. 14 (Mexico City: INAH, July 1976), Table 13; and from SHCP, “Industrialización en México,” 82; the 1844 census was compiled from Barjau Martinez et al., Tables 17–19; data for 1845 were obtained from Barjau Martinez et al., Tables 20–21; 1850 data were retrieved from Secretaria de Fomento, Colonización e Industria, Memoria que la Directión De Colonización e Industria presentó al Ministerio De Relaciones en 17 de Enero de 1852, sobre el estado de estos Ramos en el Ano Anterior (Mexico City: Secretaria de Fomento, 1852); 1854 data were obtained from Gobierno del Estado de México, Estadistica del Departamento de México (Mexico City: Fonapas, 1980)Google Scholar; 1857 data were obtained from Secretaría del Estado, Memoria de la Secretaría del Estado y del Despacho de Fomento, Colonizatión, Industria y Comercio de la República Mexicana (Mexico City, 1857); 1862 data were obtained from Hernandez, J. M. Pérez, Estadisticas de la República Mexicana (Statistics on the Mexican republic) (Guadalajara: Tip. del Go-bierno, 1862)Google Scholar; 1865 data were obtained from Fomento, Ministerio de, Memoria 1865 (Mexico City: Secretaría de Fomento, 1866)Google Scholar; 1878 data were taken from Secretaría de Hacienda, Estadísticas de la República Mexicana (Mexico City: Secretaria de Hacienda 1880); 1883 data were taken from Cubas, A. Garcia, Cuadro geográfico, estadistico, descriptivo é historico de los Estados Unidos Mexicanos (Tables on the geography, statistics, description, and the history of the United States of Mexico) (Mexico City: Secretaria de Fomento, 1884–1885)Google Scholar; 1888 data were taken from Fomento, Secretaría de, Boletin Semestral de la República Mexicana (Mexico City: Secretaría de Fomento, 1890)Google Scholar; 1891 data were taken from Cubas, A. Garcia, Mexico: Its Trade, Industries and Resources (Mexico City: Secretaría de Fomento, 1893)Google Scholar; 1893 data were taken from Dirección General de Estadística, Anuario Estadístico de la República Mexicana 1893–94 (Mexico City: Direccidn General de Estadistica 1894); 1895 data were taken from Hacienda, Secretaría de, Memoria de la Secretaría de Hacienda (Mexico City, 1896)Google Scholar; 1896 data were taken from Hacienda, Secretaría de, Estadistica de la República Mexicana (Mexico City: Secretaría de Hacienda, 1896)Google Scholar; 1898—1913 excise tax data were taken from the Ministry of Finance's periodical reports in the financial weekly Semana Mercantil (Mexico City); 1912 data were taken from “Extracto de las Manifestaciones presentadas por los fabricantes de hilados y tejidos de algodon para el semestre de enero a junio de 1912,” located at the Archivo General de la Nación (AGN), box 5, file 4 (Mexico City, n.d.); 1913 data were also located in the same archive within “Extracto de las Manifestaciones presentadas por los fabricantes de hilados y tejidos de algodón para el semestre de enero a junio de 1913,” AGN, box 31, file 2 (Mexico City, n.d.); 1914 data were taken from the July 4, 1914, issue of Mexico City's El Economista Mexicano; 1915—32 data were taken from periodical reports of excise tax data found in Secretaría de Hacienda y Crédito Público, Boletin de la Secretaria de Hacienda y Credito Público (Mexico City: SHCP, 1917–32); 1924–33 statewide data were obtained from typewritten reports of the Secretaría de Hacienda y Crédito Público—Departamento de Estadística, “Estadisticas del Ramo de Hilados y Tejidos de Algodón y de Lana,” located at the Library of the Banco de México; finally, national data for 1898—1924 were taken from Dirección General de Estadística, Anuario Estadistico 1923–1924, vol. 2 (Mexico City: Dirección General de Estadística, 1926).

44 We interpret these regressions as follows. First, to assess the importance of individual explanatory variables in these regressions, we look at their relative magnitudes and statistical significance. Second, the estimate of the intercept provides a benchmark with which to compare the coefficient estimates of the other explanatory variables because it allows us to calculate the probability of the survival of firms that do not have the characteristics of the regressors. Third, as estimated, the model is a probability function for which absolute probabilities can be calculated only if the explanatory variables X are evaluated at particular values of interest.

45 The average firm size corresponded to a 0.7 percent market share. The largest firms in the industry had market shares of 5 to 12 percent.

46 We calculated these marginal effects in two steps. First, we used specification 1 in Table 2 to calculate a probability equation as a function of size: Prob(Survival = 1 | Size) = Φ(-0.451 + 213.923 × Size + 0.445 + 0.578 - 0.543), where Φ is a normal cumulative distribution function. These are graphed in Figure 1 as “absolute probability.” Second, we measured changes in probability associated with size increments of 0.01 percent of market share. These changes correspond to the marginal effects of size on the probability of survival. The probability equation corrsponds to joint stock firms in the Central region. Varying location and firm type produced almost identical results to those shown here. To cross-check our results, we ran separate regressions with an intercept term and size as the only explanatory variables (given that estimates of the other variables in specification were neither robust nor statistically significant). We obtained similar results: marginal size effects were more important for the smallest firms and were exhausted at a size equivalent to 1.5 percent of market share.

47 Haber (fn. 29,1989), 132–34.

48 Mexico's factory owners regularly communicated these problems to the government. See Archivo General de la Nacion, Ramo del Departamento de Trabajo, Mexico City: box 45, file 3, pp. 1–5; box 52, file 12, pp. 1–3; box 90, file 17, p. 1; box 91, file 21, pp. 1–2; box 96, file 5, pp. 49–53; box 96, file 9, p. 45; box 107, file 22, p. 2; box 107, file 22, p. 2; box 110, file 28, p. 1; box 173, file 23, p. 1.

49 See Koh, Winston T. H. et al. , “Asset Pricing and Investors' Expectations: Theory and Evidence,” in Theodore Bos et al., Studies in the Financial Markets of the Pacific Basin (Greenwich, Conn., and London: JAI Press, 1994)Google Scholar. For a discussion of the value of the firm (indirectly maximized for consumers by firms) and investment costs, see Hayashi, Fumio, “Tobin's Marginal q and Average q: A Neoclassical Interpretation,” Econometrica 50 (January 1982)CrossRefGoogle Scholar.

50 Data were gathered from El Economista Mexicano and La Semana Mercantil from 1896 to 1914. Data from 1914 to 1935 were gathered from the Boletin Financieroy Minero. In cases where no shares were traded, the average of the bid and ask prices was taken. Our data set includes the four largest cotton textile producers (CIDOSA, CIVSA, CIASA, and CISAASA, which jointly accounted for roughly one-third of national output), the country's steel monopoly (Fundidora Monterrey), the largest wool textile manufacturer (the Compañia Industrial de San Ildefonso), two of the country's three most important beer brewers (the Cervecería Moctezuma and the Compañia Cervecera de Toluca y México), the country's only large-scale producer of soap and glycerin products (the Compañia Industrial Jabonera de la Laguna, which controlled 90 percent of the soap market), one of the two important producers of paper products (the Compania Industrial de San Rafael y Anexas, which monopolized the lucrative newsprint market), and two of the three firms that dominated the production of tobacco products (El Buen Tono and La Cigarrera Mexicana).

51 Jean Meyer, “Mexico: Revolution and Reconstruction in the 1920's,” in Bethell (fn. 26), 160—61.

52 Haber (fn. 29 1989), chap. 5; Moreno, Humberto Morales, “Economic Elites and Political Power in Mexico, 1898–1910,” Bulletin of Latin American Research 15, no. 1 (1996)CrossRefGoogle Scholar.

53 The fact that firms invested ahead of demand is explained by the fact that cement production tends, almost everywhere in the world, to be characterized by local monopolies. The high bulk-to-price ratio of cement means that it is economical to ship it only over short distances. In order to expand, therefore, firms must erect new production facilities in new areas of the country. And to keep out potential rivals, firms also tend to erect more productive capacity than they need in these new markets. See Johnson, Ronald N. and Parkman, Allen, “Spatial Monopoly, Non-Zero Profits, and Entry Deterrence: The Case of Cement,” Review of Economics and Statistics 65, no. 3 (1983), 431–39CrossRefGoogle Scholar.

54 Spindles constitute the most important capital input for the production of cotton textile goods, and thus the literature tends to use spindlage as the measure of capital or capacity. See, for example, Kane, Nancy F., Textiles in Transition: Technology, Wages, and Industry Relocation in the U.S. Textile Industry (Westport, Conn.: Greenwood, 1988)Google Scholar.

55 Mexican population data are from Estadisticas Históricas (fn. 27), 44.

56 Haber (fn. 29, 1989), 143–44.

57 Ibid., chap. 4; Costa and Calva (fn. 41), 198–203.

58 The four-firm ratio is calculated as the percentage of the market controlled by the four largest producers. The Herfindahl Index is calculated as the sum of the squares of the market shares of all firms in the industry. Because it uses data from all firms and because it weights firms by size, the Herfindahl Index is generally regarded as a superior index of concentration.

59 Interestingly enough, if we extend the analysis to the top ten firms and check their identity, we find that it was the same group of people who had been there prior to, during, and after the revolution. In other words, the revolution was not very revolutionary in this respect, since ownership remained concentrated in the hands of the very elite that it meant to dissolve.

60 Haber, Stephen, “Financial Markets and Industrial Development: A Comparative Study of Government Regulation, Financial Innovation, and Industrial Structure in Brazil and Mexico, 1840—1930,” in Haber, Stephen, ed., How Latin America Fell Behind: Essays on the Economic Histories of Brazil andMexico, 1800–1914 (Stanford, Calif.: Stanford University Press, 1997)Google Scholar; idem, “Industrial Concentration and the Capital Markets: A Comparative Study of Brazil, Mexico, and the United States, 1830–1930,” Journal of Economic History 51 (September 1991)Google Scholar.

61 Kane (fn. 54).

62 See Atack, Jeremy, Estimation ofEconomies of Scale in Nineteenth Century United States Manufacturing (New York: Garland Press, 1985)Google Scholar; Sokoloff, Kenneth L., “Was the Transition from the Artisanal Shop to the Nonmechanized Factory Associated with Gains in Efficiency? Evidence from the U.S. Manufacturing Censuses of 1820 and 1850,” Explorations in Economic History 21, no. 4 (1984)CrossRefGoogle Scholar; and Bernard, A. B. and Jones, C. I., “Productivity across Industries and Countries: Time Series Theory and Evidence,” Review of Economics and Statistics 78, no. 1 (1996)Google Scholar.

63 The coefficients on capital and labor from the panel regressions were normalized so that α + β = 1, because our model did not identify any economies of scale. For estimates of TFP that measure output as the real value of production, this resulted in coefficients of 0.59 for capital and 0.41 for labor. For estimates of TFP that measure output in physical units, this resulted in coefficients of 0.48 for capital and 0.52 for labor.

64 Womack (fn. 39).

65 Brown (fn. 40); Hall (fn. 40).

66 Reynolds (fn. 29), appendix C.

67 One reason that single-case studies often contribute little to theory development is that historians usually attempt to study all those aspects of the case that they judge interesting or unusual, rather than systematically study particular issues that are relevant to testing or building social science theories. The result is that social scientists are often forced to conduct their own case studies, but they do so without the detailed knowledge of past societies that is the comparative advantage of historians. See George (fn. 1), 49–51.