Bankers, Industrialists and their Cliques: Elite Networks ...



Bankers, Industrialists and their Cliques: Elite Networks in Mexico and Brazil, 1890-1915

Aldo Musacchio

Ian Read

Department of History

Stanford University

iread@stanford.edu

aldus@stanford.edu

Abstract

This paper analyzes the relationship between institutions and elite networks. We compare the institutional frameworks of Mexico and Brazil during the first stage of industrialization (1890 to 1914). During this period, what role did networks play in different institutional settings? We hypothesize that connections between firms, banks, and the government were more important in Mexico than in Brazil given their legal and political systems. Using data on interlocking board of directors and political information, we argue that Mexico relied heavily on an informal set of rules that mediated the relationship between firms, banks, and the government. The main personalities in the Mexican business network were often prominent politicians that helped firms to get charters and privileges and who facilitated the exchange of information. The state and business is seen to operate symbiotically. Brazil, on the other hand, had a relatively more standardized and open set of institutional rules and this helps explain why the Brazilian network of directors was fragmented into clusters, had weaker ties between companies, and showed low political presence. We conclude with by speculating on the theoretical relationship between institutions and networks.

Introduction

The relationship between institutions and networks has been under theorized. This paper contributes to an emerging discussion on that relationship by looking at Mexico and Brazil during their first stage of industrialization. Thus, in this paper, we compare institutional features and elite networks of these two cases. We ask what role do networks play in different institutional settings? We hypothesize that connections between firms, banks and the government were more important in Mexico than in Brazil. We find that our argument is largely confirmed by three related results. First, we show that total number of connections was slightly higher in Mexico than Brazil. Second, and more importantly, we find that there existed one dense core network in Mexico and several clusters of elite networks in Brazil. Finally, we show that politicians played an important role in the Mexican elite network.

We define networks as patterns of relationships in which two or more actors share a type of social activity over a definite period of time. A company or bank becomes part of a a network when two or more of its board members sit on the board of another company. To test our hypotheses we use a dataset of 98 Mexican firms and 242 Brazilian firms that shared board members in 1909. This creates a Mexican network of 1206 connections and a Brazilian network of 1074 connections.

We derive our hypotheses from the understanding that we have about the functions of networks within an institutional framework. We define institutions as sets of beliefs, norms and organizational features that regularize and legitimize patterns of behavior. Institutions exist in a constantly changing environment that include shifts in demography, wealth distribution, and societal beliefs (Grief, forthcoming). A particular legal framework is an institution as long as it regularize patterns of bahavior. Moreover, networks are organizational features that also can regularize and legitimize certain patterns of behavior such as the enforcement of contracts, access to credit, and the exchange of information. According to Granovetter (1985), networks are often used when certain institutional features are absent. This may include the lack of proper regulations to enforce contracts when such contracts are informally enforced through embedded relationships.

Our question is relevant for Latin American Economic history because even though the study of elites has been of primary importance in the historiography, it has rarely escaped anecdotal description. This paper contributes to the literature on elites in Latin America by providing a methodology to systematically study the networks so that researchers can better realize who were most important in the networks. Scholars working under the growing prevalence of New Economic History of Latin America have increasingly stated the connection between institutions and elite groups. But even past forms of economic and social analysis, including dependency theory, have argued that elite collusion had institutional and developmental effects. New and more formalized studies on the role of the stock market and banks in Brazil and Mexico have continued to rely on these assumptions. Unfortunately, business elite networks have not been clearly defined nor have they received a systematic analysis. Families and personalities that were part of the Mexican or Brazilian elite have been studied on an ad hoc basis, using ex-post knowledge about the elites to select them. We are not discarding the rich studies of elites in Mexico and Brazil, but only suggesting the utility of applying a new methodology to study elite networks in Latin America.

This paper, like other economic history studies (Haber 1997, Maurer 1997, Dean 1969), emphasizes that economists and policy-makers exploring 21st century issues of development in Latin American or in other industrializing regions may find part of their explanation in the institutional framework of the first stages of industrialization. The existent literature on Mexican and Brazilian economic history implicitly states that Mexican entrepreneurs continued to rely on an informal institutional structure to obtain necessary resources and information. Sources of capital, for example, were often gotten through friends and personal connections (Maurer and Sharma 2001, Gomez-Galvarriato 1999, Haber 1997, Maurer 1997). Brazilians, on the other hand, could raise larger amounts of money through other channels, like the stock market (Haber 1997, Hanley 1998).

Given that Mexico relied heavily on an informal set of rules that mediated the relationship between the economic and political; elites, firms, banks, and the government functioned symbiotically. The main personalities in the Mexican entrepreneur network were often prominent politicians that helped firms to get charters and privileges and they facilitated the exchange of information between firms by working in many boards at one time. Brazil, on the other hand, had a relatively more standardized and open set of institutional rules. This explains why the Brazilian network of directors was fragmented into clusters, had weaker ties between companies, and had showed low political presence. The final finding was somewhat surprising because the Brazilian literature on elite networks has traditionally seen politicians as playing a personal role in business (Love 1980, Wirth 1977, Levine 1979). The Mexican results, however, confirm much of what has already been stated by other economic historians, especially regarding the number of strong connections between banks and firms. In this literature, these links helped reduce information asymmetries and facilitate monitoring

Mexico and Brazil are excellent cases to test our hypothesis for two reasons. First, within Latin America, they are two of the richest countries and largest in population. GDP levels were probably similar before the 1890s and both countries relied on exports as the main source of growth during the later part of the XIX century.[1] Second, beyond historical parallels, during the period we are studying both countries went through an industrial transition at a heightened pace. By choosing two cases that share historical and some structural commonalities, we hope to maximize the numbers of variables that can be held constant

This paper is divided into four sections. In the next section we explain the methods and describe the sources used. The third section will discuss institutional differences between Mexico and Brazil. In the forth section, we present our results of the business and political elite networks while the final section concludes.

Sources and methodology

Institutional structure is determined qualitatively through informed secondary sources and systematically through a survey of Mexican and Brazilian laws on joint-stock company formation, banks, and corporate financing. The main sources of information for elite networks are taken from the Mexican Yearbook and the Brazilian Yearbook for the year 1909.[2] These books present a list of joint-stock companies, their boards of directors, object of the firm, capital, some stock prices in the previous year, and the size of the debenture issues (long-term senior secured bonds).

Network analysis, developed primarily by economic sociologists, has been used as a useful tool in exploring the economic and social relationships between firms. First, a link is established between firms that share members of boards of directors. If one or more persons work for two boards of directors of two different firms, then one or more links between the latter are created. The network is the total “web” of n number of relationships projected into a bi-dimensional space. Spatial positioning has no analytic relevance since the vertices (institutions or people) are arbitrarily positioned. These graphs serve to visualize the structure of the network and give the researcher a chance to observe patterns undetected in the data. Network analysis has a long tradition within social science and its origins can be traced to George Simmel’s sociological and philosophical work on dyads and triads at the turn of the 20th century. Recently, scholars have begun to refine computational models for network analysis and large data sets, previously too large to graph and measure, are now being used.[3]

The first step in carrying out this procedure is to create cross tabulation tables of all firms that have interlocking boards of directors. Firms are crossed with all other firms and values are derived from the number of board members shared. Next, a network visualization program called Pajek is used to visualize the networks.[4] This technique is extremely important for demonstrating specific structures of the network that are not obvious through either the cross tabulation tables or the centrality measures. For instance, cliques of firms and banks hidden in the data are revealed through their visualization.

One possible weakness of working with only one type of link between firms is that the scope of our study is limited because we ignore other informal and formal networks that may have tied entrepreneurs together. We do not have sufficient data to create networks of non-commercial ties, but we address this problem by comparing the number and importance of politicians in each set. Some kinship information exists for Brazil, but is more limited for Mexico. The differences in the quality of network data does not pose a serious problem since the limited information that we have on political and kinship ties mirror the patterns found through the data on interlocking boards.

Boards of directors of firms represent a set of shareholders that are more interested in the performance of the firm than minority shareholders. Moreover, they are more active in obtaining resources and monitoring. A close examination of the relationships between shareholders would have also shown an interesting web of links, but minority shareholders are not always compelled to work for the well being of the firm. Shareholders that have a large interest or a high stake in the profits of the firm tend to be part of the board of directors.[5] Therefore, the bias of ignoring shareholder networks is compensated by the importance of directors in managerial decisions relative to shareholders.

There is a large sociological literature that debates the functions of interlocking boards of directors.[6] We maintain that interlocking boards were used both to regulate the exchange of resources and to monitor. Information could be passed from bank to firm or government to firm that would greatly influence its later decisions. While different boards did not always serve the same function, we strongly believe that they can, at the very least, be seen as an indicator elite interests. By looking at the broad pattern of elite interests through board interlocks, one is able to recognize structural differences. This information is complemented by kinship data when available.

Institutions

To understand the different institutional arrangements in Brazil and Mexico we begin with a political history. Mexico and Brazil in the nineteenth century had two very different political trajectories and this was reflected in the legal frameworks that regulated their economic activity. Brazil had a constitutional monarchy until 1889, when this system was overthrown by a non-violent Republican revolution. Between 1889 and 1891, a provisional government was established. The new government drafted a new constitution, modified the banking laws, and enacted a comprehensive law on joint-stock companies. The Republic was democratic but had very low political participation of the population. Love (1970) has estimated that less than 10% of the population participated in elections during the period 1889-1930. Nevertheless, the federal system created through the 1891 constitution allowed far more competition of elites for representation within the federal government and competition between the states to attract business and investment. For example, states like Minas Gerais and São Paulo competed to get the best railroad network to export coffee and agricultural goods. Also, political representation of state elites alternated during the first republic. The famous café com leite arrangement occurred during this period when the republican elite of São Paulo alternated control of the presidency with the republican elite of Minas Gerais between 1891 and 1930.

Mexico followed a very different path. Mexico experienced long periods of instability after its independence from Spain. The sequence of civil wars and coups d'état did not end until 1876 when Porfirio Díaz enacted a liberal dictatorship. Before Diaz, power resided mostly in the hands of regional caciques that sharply curtailed the power of the federal government. State power during the porfiriato, however, was consolidated and centralized for 30 years, usually at the expense of the caciques (Haber, 1989). Even though both governments had similar liberal regimes (albeit less democratic in Mexico), it operated with crucial differences. Since there was such instability in Mexico during the 19th century, the rule of law was never fully established and laws were often applied arbitrarily.

The protection of property rights is an issue of relevance to compare the rule of law in both regimes. Although there are few studies on the topic, Maurer and Sharma (2001) argue that one reason for the existence of elite networks in Mexico is because of the poor protection of property rights. Since collateral was hard to obtain in case of default, banks and firms developed entrepreneurial groups to monitor their activities and enforce credit contracts. Following a similar direction, Haber, Razo and Maurer (forthcoming) argue that given the poor protection of property rights in Mexico, the government and the elites developed an implicit regulatory pact. In their notion of “vertical political integration” the government and the elites became partners in the determination of privileges and the enforcement of property rights in exchange for political support for Díaz. In Brazil, there has been little research on property rights and this prevents a fair comparison with Mexico. Bankruptcy court cases, however, have shown that the rights of creditors to their collateral was often enforced by the courts (Musacchio, 2001).

Entrance to the market was very different in Brazil and Mexico. There are two legal provisions that constrained entrance to the market. First, chartering regulations can promote or hinder entrance of new firms by setting low or high capital requirements as well as determining access according to the discretion of politicians or bureaucrats. Second, since access to capital can be an important determinant of the feasibility of new business, entering firms can be hindered by the regulations that affect the options of finance. These two aspects of access are expanded upon below.

Chartering policy in Mexico restricted access to entry. In the Commerce Code of 1884, charter approval for banks depended on the Minister of Finance. Emission banks had capital requirements between a quarter and half a million dollars while commercial banks were required to have approximately between 100,000 to 150,000 dollars. These amounts were very high for those days. The first bank that chartered in each state received a quasi-monopoly for note circulation for that state. For non-financial firms, chartering capital requirements were less onerous but government approval was still necessary. This created a problem for any who wished to create a business that would compete with several of the state protected private monopolies. For instance, companies such as dynamite factories enjoyed a monopolies granted by the government. Evidence presented by Haber (1991) also confirms that chartering laws made entrance more difficult in Mexico.

For Brazil, chartering was more of an administrative procedure. The capital required to establish a firm was 10% of capital since 1882. The approval of the charter depended only on the decision of the local Junta Comercial, the local commercial office. For banks, the approval of the charter was also relatively easy. Only banks that wanted to issue notes needed approval of the Minister of Finance.

For firm financing options, the differences between these two countries are strong and are perhaps another result of the influence of different political systems. Federalism in Brazil allowed more competition between the states to charter banks. A strong central government under Diaz helped integrate local quasi-monopolies with a system of privileges. These forces helped reduce the number of Mexico banks. The Mexican banking sector during the Porfiriato was based on two big national banks that had privileges of branching and note circulation.

Brazil also had a longer history of attempts to create a banking system through the nineteenth century and this experience certainly added to the number and strength of Brazilian banks. The richest states began chartering one or several state banks as early as the 1830s. Those efforts were not too successful until the end of the nineteenth century. By the end of the 1880s, though, the Brazilian banking system was growing in importance and there were several banks in many states. With the establishment of a federalist republic after 1891, a federalized banking system emerged that was composed of many state banks. Some of these banks had national branches, such as Banco do Brasil, the British Bank of South America Ltd., and the London and River Plate Bank Ltd. When the government legalized interstate branching in Brazil, foreign banks began to open branches with only the approval of the state government. This had a number of results that even impacted the coffee industry. Payments for Brazilian coffee exports were often received through Portuguese, Italian, French, and German banks that regularly advertised their services in the newspapers.[7]

By 1909, the Mexican banking system was composed of two big national banks, the Banco de Londres y Mexico (BLM) and Banco Nacional de Mexico (BANAMEX), and many state banks, generally one or two per state (around 40 banks total). In practice, given the prohibitive taxes on notes issued by second comer banks, only the first state bank to charter were able to successfully issue notes, which limited entry to further competition.

Brazilian firms had more formal options to get funds beyond banks. Financial markets were more developed and the federalist system promoted markets to grow in many states. For example, while in Mexico there is evidence of a fairly important stock exchange in Mexico City for mining ventures, the other stock exchange in Mexico City had very low participation of joint stock companies. The list of quotations published in the financial journals of the time shows only a handful of securities quotations. Other smaller stock exchanges operated in cities like Guadalajara, but little is known about these small exchanges. However, there was no bond market in Mexico and firms had to go to foreign markets to issue debentures. In contrast, the Rio de Janeiro Stock Exchange had approximately 100 listings throughout the period, combining shares and bonds. Furthermore, the Brazilian debenture market was buoyant and there were relatively strong stock exchanges in Rio de Janeiro, São Paulo, Recife, Salvador, and Santos.[8]

These differences caused Mexican entrepreneur to have connections to certain people in order to access large amounts of capital. The historiography on Mexico during early industrialization supports this assertion, including studies by Gomez-Galvarriato (2000), Haber (1989), Wasserman (1984), Maurer (1997), Marichal and Ludlow (1986), and Cerutti (1983, 1992). Being part of the elite network was very important to get concessions, privileges, and all sorts of political favors (Wasserman 1984, Cerutti 1983). Personal elite connections were also crucial to get equity buyers, or bank credit (Cerutti, 1983, Maurer 1997). Evidence from previous works indicates that bank credit was not a determinant source of long-term capital. For instance, manufacturing firms show very low debt-equity ratios and researchers who have studied this have emphasized the importance of the network in substituting financial markets (Haber 1989, Gomez-Galvarriato 1999). Maurer and Sharma (2001) argue that since the protection of property rights was poor in Porfirian Mexico, groups of entrepreneurs, particularly textile firms, emerged to enforce property rights through a reputation mechanism that would allow them to get credit from banks. Their study argues that credit through impersonal mechanisms was not common in Porfirian Mexico. Thus, in Mexico connections were very important not only for bank credit, but also for partners to buy equity in new ventures and for federal and state concessions or charters.

Evidence from Brazilian historiography is less broad, but generally it is argued that more impersonal sources of finance in Brazil generated less collusion between firms and entrepreneurs to get credit. For example, Haber (1991, 1997) looked closely at the textile industry and found that commercial and financial regulations in Brazil facilitated entry and access to credit. As a consequence, textile firms grew rapidly, especially when funding came through the stock exchange. Industry productivity was higher in Brazil and the number of firms increased at a greater rate. Even though banks did not play a strong role in financing the industrialization of Brazil (Triner 2000, Hanley 1995), the stock exchange were important in providing firms with finance through either equity or bond issues (Hanley 1998, Musacchio 2001).

Institutional frameworks, including political considerations, chartering and banking laws, and foreign participation, differed greatly between Mexico and Brazil. As we will show in the next section, these differences account for large dissimilarities in network structures.

Results

We hypothesize that connections between firms, banks and the government were more important in Mexico than in Brazil. By using the number of directors per firm that serve on two or more boards we develop an indicator of “connectedness.” Table 1 shows the total number of firms with board interlocks in Brazil and Mexico by sector. We can see that while board interlocks were important in both countries, they were slightly more important in Mexico. The percentage of firms sharing interlocks in both countries is high, implying that networks for both countries were important. But these numbers do not reflect the extent to which board interlocks were used by firms nor differences in the structure of the network.

Many of the sectors in both Brazil and Mexico had a very high level of connections in the aggregate. In other words, most firms in both countries used board interlocks for multiple purposes. Yet when this data is examined more closely, differences between Mexico and Brazil become apparent. Converting these figures into number of board interlocks per firm we find that connections were more important in Mexico. In Mexico each firm shared, on average, approximately two directors with other firms, while the average for Brazil was half that. Sixty percent of the Mexican firms and banks had more than one interlock, while 65% of the Brazilian firms shares only one interlock.

The figure on total number of firm interlocks shows that Mexico and Brazil both had a high number of elite interlocks, with Mexico surpassing Brazil by 12%. These numbers, however, do not reflect structural differences between the two countries and these differences become apparent when the networks are visualized. Figure 1 is a small network and it represents interlocking boards of directors for banks in Brazil. The graph shows only three dyads (two boards sharing one or more directors) with no connections between them. Very few banks shared directors in Brazil, and the few that did were not connected with each other. When we compare this to Mexican bank interlocks in figure 2 we see that there was larger and denser network for Mexico. This graph shows one large cluster, two triads, and one dyad. Large Mexican banks such as Banco Nacional de Mexico (BANAMEX), Caja de Prestamos (caja), General Bonded Warehouses of Veracuz (GBW), and Banco Central (b central) are all located in the center of the cluster. A smaller cluster of banks that are geographically distant from the large Mexico City banks is connected through bridging interlocks.

The network of interlocks between banks and manufacturing firms in Brazil is larger and more connected than the banking network (figure 3). Manufacturing firms are represented by dark vectors, while banks are white. This graph demonstrates some of the uniqueness of the overall Brazilian elite network: a fairly large cluster of Rio de Janeiro and São Paulo firms that is unconnected to independent small clusters. Many of these clusters represent firms from other states such as Rio Grande do Sul, Maranhão and Bahia. For example, Progresso Industrial, a firm that produced textile manufactures in Rio Grande do Sul, is connected to two banks in that same state. Interestingly, banks do not play a central role in these clusters, perhaps because, as some have claimed, banks did not participate actively in the funding of industrialization of Brazil during this period (Triner 2000, Hanley 1995).

Figure 4 shows Mexican bank-manufacturer interlocks. The core cluster becomes apparent again as many of the manufacturing firms (dark vectors) were strongly connected to the core cluster. Of the 12 manufacturing and textile firms represented, eight are closely tied to many banks within this cluster. Only fours firms are located in the periphery of the cluster. The overall elite network differs from Brazil in that there is a core pattern of elite relationships that repeats itself in every sector. This is evident by comparing the bank-bank network (figure 1) with the bank-manufacture network (figure 4). While positioning in arbitrary with these methods, we can still see that there exists a core structure of elites that is connected, through Banco Central (b central) to a smaller cluster of banks and manufacturers that include Banco Minero and the Compañía Industrial la Laguna. The triad between Banco de Querétaro, the Banco Internacional Hipotecario and the Banco de Hidalgo that was found in the bank network is now connected to the core through the cigarette manufacturer El Buen Tono. This demonstrates that firms and banks were connected to the core on varying levels, and small groups or individual firms that was not connected to the core in one sector might be connected to this core in another sector.

The bank-manufacturer network in Mexico is also interesting because it contributes to the current debate on the sources of finance of the Mexican industry during this period. The strong ties between manufacturers and banks shows that interlocks might have played a very important role both in the ability of firms to get credit through insider lending and the capacity of banks to monitor firm activities. The results already studied for textiles in Maurer and Sharma (2001) and for other firms in Maurer (1997) can most likely be extended to the whole manufacturing sector.

Another way of representing the networks is through connections between the directors themselves. While this may give less sense of the difference between sectors of the economies, it does show who were central figures in these networks. Figures 5 and 6 show elite networks connected by directors. For sake of clarity, we have excluded names from these graphs to emphasize structural differences. These two graphs show the striking differences in network structure between Mexico and Brazil. The Brazil graph repeats what was found with connections between firms: a number of unconnected clusters. The Mexican elite network, on the other hand, has a dense core with only a few separate dyads. The small cluster protruding from the core is a group of English financiers who invested heavily in Mexican railroads and mines.

One way to make these graphs manageable is to look at the network of directors that had three or more connections. Figures 7 and 8 show the patterns of relationships between the top businessmen, bankers and politicians in Brazil and Mexico. These networks, while smaller in size, still reflect the same clustered patterns for Brazil and core network for Mexico. Looking closely at the individuals that composed these networks, we find a few surprises. As seen in figure 7, there are far more politicians in the Mexican core elite than in the Brazilian network and these results are overwhelming. In table 2 we present the rank of Mexican personalities according to their centrality.[9]

The last column is used as the main indicator of centrality because it is a more robust measure than other measures. Eigenvalue recursively takes into account the number of connections of a personality and the connections of those to whom that personality is connected to, repeating the procedure throughout the whole network. According to this rank we found political information for 1909 and the 30 years prior.

Additionally, this table shows that the most connected Mexican directors were also important politicians. Many of them were congressmen and top authorities in Mexico City. In fact, most of these personalities were important congressmen that had participated in drafting important laws, such as the banking law of 1897. Some, like Pablo Macedo and Joaquín Casasús, were lawyers who could offer legal advice to the firms that they represented. Others were the financial representatives of banks and companies who could offer financial information, served as monitors for the banks, and offered connections to the firms for credit. For example, the Banco de Londres y México and the Banco Nacional de México (Banamex) were the only banks with the right to issue notes that were used as legal tender nationwide. Banco de Londres y Mexico was legally represented by Joaquín Casasús, who fought a stiff political and legal battle against Banamex to win this privilege. (Ludlow and Salmerón 1997). Casasús and other politicians attended the Monetary Commission which, in 1905, drafted the Monetary Law placing Mexico on the gold standard.

Friendship and kinship ties were also important in determining the structure of the Mexican network. For instance, Pablo Macedo and Hugo Scherer (Jr.) were partners of the Minister of Finance, Jose Yves Limantour, in several businesses. The Limantour family had important interests in many of the companies that appear at the core of our network such as the Banco de Londres y México, the San Rafael Paper Company (an industrial monopoly), the cigarette manufacturer El Buen Tono, and others. Other directors who were part of the boards of these companies were also close friends of the Minister of Finance. For instance, Pablo Macedo directed a newspaper sponsored by Limantour and Hugo Scherer Jr. and Julio Limantour, brother of Jose Yves Limantour, were partners in an investment bank in Mexico City (Ludlow and Salmerón 1997). Perhaps the best example of the union of politics and business was Porfirio Díaz Jr, the son of the long-ruling dictator Porfirio Díaz. Diaz Jr, served in the boards of many important companies, including two banks, El Buen Tono, the Mexican Eagle Oil Company, one railroad company, and the biggest utility company in Mexico. Interestingly, Díaz Jr. also sat on the board of the Banco Internacional e Hipotecario de México with Minister of Finance Limantour. The last of several anecdotal examples that can be taken from this list is the vice Minister of Finance Roberto Nuñez who appears with Diaz Jr on the board of El Buen Tono. Nuñez worked for Banco Nacional de Mexico and Caja de Préstamos, two of the biggest financial institutions at this time.

The role of these politicians in the elite network was very important since information and concessions or privileges were passed mostly from politicians to the companies for which they represented. Without these top political figures, economic survival could have been quite difficult during Porfirian Mexico. Thirty seven percent of the total firms were represented by at least one of the top 15 directors included in table 2. This group of top political elite also worked for firms that controlled 50% of the total capital of joint stock companies in Mexico. If the foreign firms are excluded, this figure would be much higher. Interestingly, the top political elite had the largest presence in banking and manufacturing. In banking they controlled 35% of banks, accounting for 65% of total commercial and mortgage bank capital in Mexico. While in manufactures they controlled nine of the 12 firms in our database, accounting for 80% of total manufacturing capital.

The top political elite also show their importance in personally influencing business when we look at foreign firms. Many of these firms, like railroad companies, were required by law to have an overseer board in Mexico. A majority of foreign firms (chartered and managed abroad) also included these top personalities in their overseeing boards in Mexico, likely giving them access to lower the costs in dealing with the government. In turn, the Mexican political elites were able to control important business and financial information, even when they were originally foreign-owned. The Mexican Eagle Oil Company exemplifies this arrangement. It was incorporated in Mexico by Sir Weetman D. Pearson after he discovered oil while building a railroad to across Mexico. He later sold the company to a new group of entrepreneurs but maintained a stake in the company. A new board was formed that included Guillermo de Landa y Escandón, John B. Body, Enrique Creel, Porfirio Díaz Jr., Luis Elguero, Pablo Macedo, Fernando Pimentel y Fagoga, Luis Riba, Enrique Tron and Pearson. All of these directors are in the top 15 list we present, while Pearson missed the list my one rank. The Mexican Eagle Oil Company later became one of the two largest oil companies in Mexico.

Turning now to Brazil, we see far less evidence of political participation. By comparing our list of directors with comprehensive lists of all federal and state delegates during this period, we found only three directors that had political ties. First, Antonio Maia (coded as “maia”) served on the Congress and as Secretary of Agriculture in 1892. Manuel Py (py) served on the 2nd National Legislature (1892-1896) for Rio Grande de Sul. Possidonio Manso da Cunha, Jr also served on for this state on the 3rd National Legislature (1897-1899). Overall, we found very few politicians represented in the board of directors.[10]

Unlike Mexico, some work has been done on political and kinship networks for Brazil using our methodology. Joseph Love (1980) looked at connections between these elites for the state of São Paulo. Using his data, we created a political and kinship network that includes director members in figure 9. Two interesting facts emerge from this network. First, none of the directors (dark vectors) are included in our top business elite lists. Second, since the same clustered patterns repeats itself in the this network, we can assume that when business, political and kinship networks are layered, they will shared structural similarities. This is no surprise since business and political elites moved in the same social circles. Yet these elite circles seemed to have been regionally clustered in Brazil. We can imagine that if there existed a political and kinship network for Mexico, it would probably resemble the same core cluster pattern that we have already seen.

In sum, our results have shown: 1) that networks played a slightly more important role for Mexico; 2) that the structure of these networks were important for understanding how they might have been used; and 3) that politicians played a stronger role in the Mexican networks and were largely absent for Brazil.

Conclusion

This paper has demonstrated that given an institutional framework, it might be possible to predict the structure of elite networks. This prediction has rested on the assumption that people will use networks less for two reasons. First, they might use informal relationships less to access capital when this resource can be raised through formal mechanisms such as the stock market. Second, when informal monitoring and enforcement is superceded by a relatively strong rule of law networks have less utility. Beyond confirming our basic prediction, we were surprised by additional results. Politicians were numerous in the Mexican elite network, but uncommon in the Brazilian network. That finding fits nicely with the state/business co-dependence model that has been theorized in the historiography of Mexico but contradicts past interpretations of Brazil’s political economy. The other surprise was the degree to which Mexican elite personalities were found to have participated in a core pattern of relationships and that this pattern repeated itself in the network of every sector of the economy.

We did not attempt to establish lines of causality between networks and institutions. Since networks can also regularize patterns of behavior, they can be a strong part of the institutional framework. We found this to be the case in Mexico. Since networks play a weak or strong role in informally enforcing contracts, they are more a part of institutions than a consequence. In addition, institutional frameworks that support strong informal interaction via networks often benefit those who are in a position to rewrite laws. Since Mexican lawmakers were often those who were most embedded in the networks and who profited most from those networks, they designed laws that perpetuated the institutional structure. Interestingly, the Mexican Revolution was largely initiated by Francisco Madero whose family is only distantly connected to the core group of elite in our 1909 data (see figure 7).

When viewing the banks and industrial firms in Mexico and Brazil as organizations, we can shed some light on the impact of networks. Since organizations have to constantly adapt themselves to their changing environment, they must access various resources to make these changes. Besides physical resources such as capital, information is necessary to monitor the activities of other actors. The network undoubtedly affects the access to both capital and the manner which monitoring is performed. Organizations and people that are embedded in strong dense networks will more likely receive information and capital from sources within this network, while organizations that are outside of networks rely more on impersonal sources of information and capital. Mexican firms, for example, relied to a great deal on the capital from the banks that were close to them in their networks. In contrast, Brazilian firms could better issue bonds or raise money through the stock market and during this period of industrialization larger sources of capital came from other sources besides banks. Brazilian firms grew larger and faster because larger sources of capital were more readily available.

One aspect of the relationship between networks and institutions is how firms interact with banks and the state to access to capital. While networks did more than provide easy, personal loans between firms and banks, the prevalence of banks in the Mexican networks raises interesting questions. As this paper has described, Mexico had greater restrictions in the financial system. Barriers to entry were greater and there were fewer options of long-term finance for domestic firms. Since foreign firms in Mexico had access to financial markets in London, Paris, and New York, they were able undertake enormous investments such as railroads, utility companies, and refining plants. Yet those that were unable to cultivate relationships with the dense core of domestic banker and industrialist elite probably met with less success. In Brazil, we found that elite networks were relatively diffuse, connections to banks were not an important part of the networks, and there is no strong core of connections in any sector. Foreign firms also played an extremely important role in Brazil but were not strongly embedded in any core network of elite. Many of these foreign firms and domestic firms outside of the networks also issued long-term debentures.

A dense group like the Mexican core elite might not be economically harmful if new entrants have other options for getting credit. The Mexican elite, however, monopolized access to the market, restricted credit to its clique, and designed policies and laws to perpetuate the system. The collusion between the government and the elites has been a common facet of Mexican historiography.[11] This union of business and political elite in Mexico and their relative separation in Brazil has wide ranging consequences. The Mexican state and businesses internalized a co-dependent relationship and being part of the network became necessary for large businesses to access credit and information domestically. Moreover, large firms relied on these connections to have access to further concessions and privileges from the government. While firms actually might have found it profitable to operate under these informal rules in the short run, in the long run, this probably had negative consequences for the economy. An economic system that has high costs of entry, where access to credit depends on personal relations, and is dominated by a clique, negative implications on future income distribution and overall economic growth is easily imaginable. Such an informal institutional model might not be detrimental when economies are small since property rights can be enforced and monitoring can be performed through the network. Economic growth, however, can be constrained when business can only grow within the regulation of a core group of elite.

In this paper, we have speculated on the relationship between the institutional frameworks in early industrial Mexico and Brazil and the business and political elite networks. While the evidence we presented contributes to current debates in the economic historiography of Latin America, we hope that some of these ideas can extrapolated onto a theoretical level. This paper is a small step toward a more general theory of the relationship between institutions and networks.

Bibliography

Almanak Administrativo, Mercantil e Industrial ("Almanak Laemmert"). 1889-1914. Rio de Janeiro.

Brazilian Year Book 1909. London: McCorquodale & Co.

Jornal Do Commercio.1889.

Mexican Year Book . A Statistical, Financial, and Economic Annual, Compiled from Official and Other Returns. 1909-1910. 1910. London: McCorquodale & Co.

Brazil. Congreso. Deputados Brasileiros 1826-1976. 1976.

Cano, Wilson. 1987. Raízes da Concentração Industrial em São Paulo. São Paulo: T. A. Queiroz Editora.

Cardoso, Fernando Henrique, and Enzo Faletto. 1979. Dependency and Development in Latin America. Berkeley: University of California Press.

Cerutti, Mario. 1992. Burgesía, capitales e industria en el norte de México. Mexico: Alianza and Universidad Autonoma de Nuevo Leon.

———. 1983. Burguesía y capitalismo en Monterrey. México: Claves Latinoamericanas.

Chirino, Jorge. 1999. "Pozos, de coyotes, crac, y optimismo: origen y clausura de la Bolsa de México, 1895-1896." B.A. Thesis, Instituto Tecnológico Autónomo de México.

Coatsworth, John and Alan Taylor, Eds. 1998. Latin America and the World Economy Since 1800. Cambridge, Mass: DRCLAS/Harvard University Press.

Dean, Warren. 1969. The industrialization of São Paulo, 1880-1945. Austin: The University of Texas Press.

Gomez-Galvarriato. 1999. The Impact of the Revolution: Business and Labor in the Mexican Testile Industry, Orizaba, Veracruz, 1900-1930. Harvard University: Ph D dissertation.

Granovetter, Mark. 1994. "Business Groups." inIn Handbook of Economic Sociology. Ed N. Smelser, and R. Swedberg, 453-75. Princeton: Princeton University Press.

———. 1985. "Economic Action, Social Structure, and Embeddedness." In American Journal of Sociology 91, no. 3: 481-510.

Greif, Avner. 1994. Cultural Beliefs and the Organization of Society: A Historical and Theoretical Reflection on Collectivist and Individualist Societies. Journal of Politica Economy 102, no. 5: 912-50.

———. forthcoming. Historical Institutional Analysis. Cambridge and New York: Cambridge University Press.

Haber, Stephen. 1998. "The Efficiency Consequences of Institutional Change: Financial Market Regulation and Industrial Productivity Growth in Brazil, 1866-1934." Latin America and the World Economy Since 1800. Eds. John and Alan Taylor Coatsworth, 275-322. Cambridge, Mass.: DRCLAS and Harvard University Press.

———, Ed. 1997. How Latin America Fell Behind. Stanford: Stanford University Press.

———. 1991. Industrial Concentration and the Capital Markets: A Comparative Study of Brazil, Mexico and the United States, 1830-1930. Journal of Economic History 51, no. September: 559-80.

———. 1989. Industry and Underdevelopment. Stanford: Stanford University Press.

Haber, Stephen, Armando Razo, and Noel Maurer. 2000. The Political Economy of Instability: Political Institutions and Economic Growth in Revolutionary Mexico.

Hanley, Ann. 1998. "Business Finance and the São Paulo Bolsa, 1886-1917." Latin America and the World Economy Since 1800. Ed. John and Alan Taylor Coatsworth, 115-38. Cambridge, Mass.: DRCLAS and Harvard University Press.

———. "Capital Markets in the Coffee Economy: Financial Institutions and Economic Change in Sao Paulo, Brazil, 1850-1905." Stanford.

Hewlett, S. Ann, and Richard S. Weinert. 1982. Brazil and Mexico: patterns in late development. Philadelphia: Institute for the Study of Human Issues.

Lamoreaux, Naomi. 1994. Insider Lending. Cambridge: Cambridge University Press.

Lang JR, Lockhart DE. "Increased environmental uncertainty and changes in board linkage patterns." In Aca .

Leff, Nathaniel. 1982. Economic Structure and Change, 1822-1947. London: George Allen & Unwin.

Levine, Robert M. 1978. Pernambuco in the Brazilian federation, 1889-1937. Stanford, Calif: Stanford University Press.

Levy, Maria Barbara. 1977. Historia da Bolsa de Valores do Rio de Janeiro. Rio de Janeiro: IBEMEC.

Love, Joseph LeRoy. 1970. Political Participation in Brazil, 1881-1969. Luso-Brazilian Review 7, no. 2: 3-24.

———. 1980. São Paulo in the Brazilian Federation, 1889-1937. Stanford, Calif: Stanford University Press.

Ludlow, Leonor, and Alicia Salmerón. 1997. La emisión de papel moneda en México: Una larga negociación político-financiera. México: SHCP/Talleres de Impresión de Estampillas y Valores.

Ludlow, Leonor, and Carlos Marichal. 1986. Banca y poder en México, 1800-1925. 1a ed ed. Enlace. Historia: Colección Enlace. México: Grijalbo.

Mace, ML. 1971. Directors: Myth and Reality. Boston: Harvard Business School.

Marichal, Carlos, and Mario Cerutti. 1997. Historia de las grande empresas en México, 1850-1930. 1. ed ed. Obras De Economía Latinoamericana: Economía Latinoamericana (Mexico City, Mexico). Nuevo León, México: Universidad Autónoma de Nuevo León. Fondo de Cultura Económica.

Maurer, Noel. "Finance and Oligarchy: Banks, Politics, and Economic Growth in Mexico, 1876-1928." Stanford University.

Maurer, Noel, and Tridib Sharma. 2001. Enforcing Property Rights Through Reputation: Mexico's Early Industrialization, 1878-1913. Journal of Economic History 61, no. 4.

Mizruchi, MS. 1982. The American Corporate Network: 1904-1974. Beverly Hills: Sage.

Musacchio, Aldo. "Institutions and Modernization: The Rio de Janeiro Stock Exchange and the Industrialization of Brazil, 1889-1930". Presented at the Economic History Association 61st Annual Meeting.

Musacchio, Humberto. 1999. Milenios de México. México: Raya en el Agua.

North, Douglass. 1990. Institutions, Institutional Change, and Economic Performance. Cambridge: Cambridge University Press.

Pelaez, Carlos Manuel and Wilson Suzigan. 1976. Historia Monetaria do Brasil. Brasilia: Universidade de Brasilia.

Pennings JM. 1980. Interlocking Directorates. San Francisco: Jossey-Bass.

Riberiro de Andrada, Carlos. 1923. Bancos de emissão no Brasil. Rio de Janeiro: Leite de Ribeiro.

Scott, John. 1991. Social Network Analysis: a handbook. London: Sage.

Scott WR. 1992. Organizations: Rational, Natural, and Open Systems. Engrewood Cliffs, NJ: Prentice Hall.

Selznick, P. 1957. Leadership in Administration. New York: Harper & Row.

Smith, Peter, producer. 1999. "Political Elites in Mexico, 1900-1971." Web page. Available at .

Suzigan, Wilson. 2000. Industria Brasileira: origem e desenvolvimento. São Paulo: Hucitec, Ed. da Unicamp.

Swedberg, Richard, Ed. 1999. Essays in Economic Sociology. Princeton: Princeton University Press.

———. 1998. Max Weber and the idea of economic sociology. Princeton: Princeton University Press.

Triner, Gail. 2000. Banking and Economic Development: Brazil, 1889-1930. New York: Palgrave.

Wasserman, Mark. 1992. "La inversión extranjera en México, 1876-1910: un estudio de caso del papel de las élites regionales." Historia Económica de México. Comp. Enrique Cárdenas, 267-89. Vol. 3. México: Fondo de Cultura Económica.

Wasserman, Stanley. 1997. Social Network Analysis: methods and applications. Cambridge and New York: Cambridge University Press.

Wirth, John D. 1977. Minas Gerais in the Brazilian federation, 1889-1937. Stanford, Calif: Stanford University Press.

Zeitlin, M. 1974. Corporate ownership and control: the large corporation and the capitalist class. American Journal of Sociology 79: 1073-119.

Tables and Figures

Table 1. Firms with board interlocks in Mexico and Brazil by Sector

|  |  |MEXICO |  |  |BRAZIL |  |

|  |Total number |Firms sharing |as a % |Total number |Firms sharing |as a % |

| |of firms |directors | |of firms |directors | |

|AGR | | | |13 |7 |54% |

|BAN |34 |28 |82% |43 |30 |70% |

|COFF | | | |7 |6 |86% |

|CONS | | | |8 |8 | |

|FOOD |1 |1 |100% |9 |5 |56% |

|IE | | | |6 |1 |17% |

|INS | | | |34 |24 |71% |

|K |1 |1 |100% |5 |2 |40% |

|MAN |12 |12 |100% |99 |72 |73% |

|MIN |34 |23 |68% |31 |14 |45% |

|NAV | | | |12 |10 |83% |

|PORT |1 |1 |100% |9 |7 |78% |

|RR |20 |18 |90% |31 |23 |74% |

|RUBBER | | | |9 |4 |44% |

|SVC | | | |10 |3 |30% |

|TEL |2 |1 |50% |4 |2 |50% |

|UT |8 |6 |75% |40 |25 |63% |

|Oil |3 | |0% | | | |

|Total |116 |91 |78% |370 |243 |66% |

Note: AGR=Agriculture; BAN=Banks; COFF=Coffee Companies; CONS= Construction firms; FOOD= Meat Packing, food processing, or sugar mills; IE=Import/Export Business; INS=Insurance companies; K=Capital manufacturers, for Mexico is a steel mill; MAN=Manufacturing in general: textiles, cigarretes, and other goods; MIN=Mining companies; NAV=Shipping companies; PORT= Port operating and construction companies; RR=Railroad companies; RUBBER= Rubber export companies; SVC= Services, media, and printing companies; TEL= Telephone and telegraph companies; UT=Utilities: Tramways, electric, and water companies; Oil=Oil companies.

Table 2. Centrality measures for the most important Mexican directors

|Rank |Last name |First name |Degree |Closeness |Betweenness |Eigenvector | |

|1 |Macedo |Pablo |67.4 |74.2 |12.4 |39.9 | |

|2 |Landa y Escandón |Guillermo |65.2 |71.9 |11.1 |38.8 | |

|3 |Scherer |Hugo |60.9 |71.9 |16.0 |35.1 | |

|4 |Brown |E N |54.3 |65.7 |5.2 |33.7 | |

|5 |Elguero |Luis |47.8 |63.9 |2.3 |33.0 | |

|6 |Pimentel y Fagoga |Fernando |52.2 |66.7 |7.0 |32.5 | |

|7 |Signoret |José |45.7 |63.0 |2.7 |30.6 | |

|8 |Creel |Enrique |45.7 |61.3 |3.7 |29.4 | |

|9 |Riba |Luis |41.3 |59.7 |1.7 |28.8 | |

|10 |Body |J B |39.1 |60.5 |2.8 |27.2 | |

|11 |Casasús |Carlos |41.3 |61.3 |3.9 |26.6 | |

|12 |Tron |Henri |37.0 |60.5 |1.6 |25.7 | |

|13 |Díaz |Porfirio, Jr. |34.8 |59.0 |1.0 |25.6 | |

|14 |Casasús |Joaquín |32.6 |58.2 |1.4 |23.8 | |

|15 |Nuñez |Roberto |30.4 |56.8 |0.5 |21.7 | |

|16 |Pearson |Witman |28.3 |54.8 |1.4 |19.7 | |

|17 |Dondé |Manuel |26.1 |55.4 |0.4 |19.6 | |

|18 |Escandón |Pablo |28.3 |56.1 |1.1 |19.0 | |

|19 |Pugibet |Ernesto |26.1 |54.8 |0.3 |18.8 | |

|20 |Barroso |Luis |32.6 |58.2 |2.5 |18.7 | |

|21 |Otto |Ernesto? |23.9 |54.8 |0.1 |18.7 | |

|22 |Alcazar |Ramon |28.3 |56.1 |0.4 |18.4 | |

|23 |Araoz |Manuel |26.1 |55.4 |0.3 |17.5 | |

|24 |Brown |Robert C |30.4 |54.1 |1.1 |17.4 | |

|25 |Honey |Ricardo |21.7 |52.9 |0.3 |17.2 | |

|26 |Ramos |José S. |21.7 |53.5 |0.2 |16.2 | |

|27 |Ebrard |Juan |23.9 |50.0 |0.4 |15.5 | |

|28 |Lemmens |J J |23.9 |54.8 |0.3 |15.4 | |

|29 |Camacho |Sebastian? |19.6 |52.9 |0.0 |15.0 | |

|30 |Wood |E. R. |21.7 |51.1 |0.1 |13.3 | |

|31 |Pearson |F.S. |21.7 |51.1 |0.1 |13.3 | |

|32 |Lash |Miller |21.7 |51.1 |0.1 |13.3 | |

|33 |Horne |Sir. William van |21.7 |51.1 |0.1 |13.3 | |

|34 |Gow |Walter |21.7 |51.1 |0.1 |13.3 | |

|35 |Pardo |José Maria |17.4 |52.9 |0.0 |12.2 | |

|36 |Schroeder |Ernesto |17.4 |49.5 |0.0 |11.6 | |

|37 |Lash |Z. A. |19.6 |47.4 |0.3 |9.9 | |

|38 |Tárnava |Constantino de |13.0 |52.3 |4.6 |8.2 |Bridge |

|39 |Mancera |Tomás |10.9 |47.9 |0.3 |7.8 | |

|40 |Mackew |Dr. S. |10.9 |46.5 |0.0 |7.3 | |

|41 |Sobey |R. T. |8.7 |48.4 |0.0 |7.0 | |

|42 |Hartman |Eduardo |6.5 |45.1 |0.0 |5.1 | |

|43 |Hoghton |G. W. |6.5 |42.2 |0.0 |4.4 | |

|44 |Cantú Treviño |Manuel |4.3 |41.4 |0.0 |3.1 | |

|45 |Madero |Ernesto |8.7 |44.2 |4.6 |2.6 |Bridge |

|46 |Garza |Marcelino |6.5 |35.4 |0.4 |0.7 | |

|47 |Terrazas |Juan |4.3 |31.1 |0.0 |0.2 | |

Note: For methodology see below**.

Table 3. Porfirian Personalities:

|Rank |Last name |First name |Political profile |

|1 |Macedo |Pablo |Important Lawyer and friend of Minister of Finance. Congressman 4 times between 1880 and 1911. Many times president of Congress, president of the|

| | | |commission of finance and government budget.. |

|2 |Landa y Escandón |Guillermo |Senator since 1878, and Major (1900-03) and governor (1903-1911) of Mexico City. Member of the Monetary Commission (1903-05). |

|3 |Scherer |Hugo |Representatives of important European mining and banking companies. Intermediary of the government and London Banks. Member of the Commission |

| | | |that drafted the banking law (1896). Member of the Monetary Commission (1903-05). |

|4 |Brown |Ernesto | |

|5 |Elguero |Luis |Major of Mexico City and advisor to the government (Lawyer). |

|6 |Pimentel y Fagoga |Fernando |Major of Mexico City (1903-1911), member of the Monetary Commission (1903-05). |

|7 |Signoret |Jose | |

|8 |Creel |Enrique |4 times state congressman for Chihuahua and 3 times national congressman for that state. Substitute Governor of Chihuahua (1903-1906) and later |

| | | |governor (1906-1911); ambassador to the US in 1906-1907; Minister of Foreign Relations (1910-1911). President of the Mexican Bankers Association |

| | | |and of the Mexican Society for Geography and Statistics. Married the daughter of Chihuahua's influential Luis Terrazas (Governor, Senator, and |

| | | |Cacique). Monetary Commission (1903-05) |

|9 |Riba |Luis | |

|10 |Body |J B | |

|11 |Casasus |Carlos |Brother of Joaquin Casasús (below). |

|12 |Tron |Henri | |

|13 |Díaz Jr. |Porfirio |Son of dictator Porfirio Diaz |

|14 |Casasús |Joaquín |Congressman (1886-1907) and Senator (1907-1911). Active lawyer, leader of the congressional commission in charge of drafting the code of commerce|

| | | |1889 and the banking law of 1897. Also worked for the Monetary Commission (1903-05). Ambassador of Mexico in the US (1905-1906). |

|15 |Nuñez |Roberto |Vice-minister of Finance (Lawyer) |

Sources: Musacchio (1999), Haber (1986), Ludlow and Salmeron (1997), and Peter Smith [producer], Political Elites in Mexico, 1900-1971.

Note: for methodology of how we determined the ranking of each personality see below**

|***Brief note on the methodology for tables 2 and 3. | |Table 4. Index of connectedness to banks (firms connected to banks as % of total firms) |

|To determine the rank of importance among Mexican | |  |

|directors we studied the centrality of the figures | |  |

|that worked for 3 firms or more. For tables 2 and 3 | |MEXICO |

|the rank was obtained by sorting the centrality of | |  |

|company directors according to their eigenvalue | |  |

|measure of centrality. The measures of centrality used| |BRAZIL |

|are: 1)Degree, which measures centrality by counting | |  |

|the number of connections/board interlocks that a | | |

|point has with other points. 2)Closeness, which | |  |

|measures the distance of a point to all other points | |Total number of firms |

|of the network, the more central the figure is, the | |Firms connected to banks |

|less distanced it should be from all the points in the| |as a % |

|network. 3)Betweenness, which measures the degree to | |Total number of firms |

|which a point links two groups of people, it takes | |Firms connected to banks |

|into account the number of points that are connected | |as a % |

|through the existence of that observation; and 4) | | |

|eigenvalue, which measures recursively the connections| |AGR |

|of a point by looking at the connections of those | | |

|points connected to it and the connections of the | | |

|points that subsequently get connected to it. All of | | |

|our measures of centrality are normalized by the | |13 |

|number of total connections existent in the network | |1 |

|according to each measure and then indexed so that the| |8% |

|maximum possible centrality is equal to 100. | | |

| | |BAN |

| | |34 |

| | |26 |

| | |76% |

| | |43 |

| | |6 |

| | |14% |

| | | |

| | |COFF |

| | | |

| | | |

| | | |

| | |7 |

| | |5 |

| | |71% |

| | | |

| | |CONS |

| | | |

| | | |

| | | |

| | |8 |

| | | |

| | | |

| | | |

| | |FOOD |

| | |1 |

| | |1 |

| | |100% |

| | |9 |

| | |3 |

| | |33% |

| | | |

| | |IE |

| | | |

| | | |

| | | |

| | |6 |

| | |1 |

| | |17% |

| | | |

| | |INS |

| | | |

| | | |

| | | |

| | |34 |

| | |17 |

| | |50% |

| | | |

| | |K |

| | |1 |

| | |1 |

| | |100% |

| | |5 |

| | |2 |

| | |40% |

| | | |

| | |MAN |

| | |12 |

| | |11 |

| | |92% |

| | |99 |

| | |27 |

| | |27% |

| | | |

| | |MIN |

| | |34 |

| | |9 |

| | |26% |

| | |31 |

| | |1 |

| | |3% |

| | | |

| | |NAV |

| | | |

| | | |

| | | |

| | |12 |

| | |3 |

| | |25% |

| | | |

| | |PORT |

| | |1 |

| | |1 |

| | |100% |

| | |9 |

| | |3 |

| | |33% |

| | | |

| | |RR |

| | |20 |

| | |11 |

| | |55% |

| | |31 |

| | |6 |

| | |19% |

| | | |

| | |RUBBER |

| | | |

| | | |

| | | |

| | |9 |

| | |1 |

| | |11% |

| | | |

| | |SVC |

| | | |

| | | |

| | | |

| | |10 |

| | | |

| | |0% |

| | | |

| | |TEL |

| | |2 |

| | |1 |

| | |50% |

| | |4 |

| | |1 |

| | |25% |

| | | |

| | |UT |

| | |8 |

| | |4 |

| | |50% |

| | |40 |

| | |12 |

| | |30% |

| | | |

| | |Oil |

| | |3 |

| | |1 |

| | |33% |

| | | |

| | | |

| | | |

| | | |

| | |Total |

| | |116 |

| | |66 |

| | |57% |

| | |370 |

| | |89 |

| | |24% |

| | | |

| | |Note: Each connection represents a board interlock between a firm and a bank. |

Table 5. Connections to banks per sector, Brazil vs. Mexico

|  |  |

| | |

[pic]

[pic]

[pic]

[pic]

[pic]

[pic]

[pic]

[pic]

-----------------------

[1] GDP levels are difficult to determine from this period but other indicators, such as the number mileage of telegraph lines, indicate similarities in levels of development between the two countries.

[2] Unfortunately, the English who wrote the Brazilian Yearbook did not publish any earlier or subsequent yearbooks. The Mexican Yearbook has a wider length of publications and one interesting extension of this project would be to add both a temporal component and to include data from the 1913 yearbook. In that year, many of the personalities who figure prominently in our networks for 1909 are gone due to the Mexican Revolution. Many, such as Porfirio Diaz, Jr., had taken up comfortable exile in Paris.

[3] See Wasserman and Faust (1994) for a broad survey of network methods and theoretical applications.

[4] Fortunately, Pajek is available on the internet at:

[5] We are implicitly assuming that shareholders elect board members by maximizing an expected profits function, using the connections of the board members as one important independent variable. Thus the portfolio of members of the board tries to maximize connections in other sectors of the economy, in the financial system, and in the political arena.

[6] Sociologists have long debated the function of interlocking board of directos. Interlocks in this literature can be used for: 1) collusion (Mizruchi 1982, Pennings 1980); 2) cooptation and monitoring (Land and Lockhart 1990, Sheard 1993); 3) legitimacy (Selznick 1957, Scott 1992); 4) career advancement (Mace 1971); or 5) social cohesion (Zeitlin 1976). We take the position that board interlocks were important for exhanging resources and information between two organizations. These networks certainly led to greater social cohesion.

[7] The number of banks in each country is difficult to trace throughout the nineteenth century, for Brazil see Pelaez and Suzigan (1976), Triner (2000), and 19th century Almanak Laemmert. This later source shows that in Brazil by 1909 there were at least one or two state banks in most Brazilian states, as well as a system of national branches for many Rio de Janeiro banks. In Mexico we base our accounts on Maurer (1997). For advertisements of foreign banks in Brazilian newspapers see Journal do Commercio, several years.

[8] For Mexico see Maurer (1997) and Chirino (1999), for the Brazilian stock markets we base our arguments in Hanley (1998), Haber (1998), and Musacchio (2001).

[9] The simplest way to measure centrality is by looking at the degree measure, which determines the centrality of a point in terms of the number of connections/board interlocks that it has with other points. This is a very “local” measure of centrality since it does not take into account if someone is connected to very important people, i.e. people with many connections. Thus there is a measure of global centrality that measures the distance of a point to all other points of the network called “closeness”. The more central the figure is, the less distanced it should be from all the points in the network. But sometimes points link two groups of people and in that sense they are more central, thus there is a measure of betweenness that takes into account the number of points that are connected through the existence of that observation that is in between. For our study this measure is important because we have regional elite networks that sometimes are connected to the core network or to other networks through very important personalities or firms. Those personalities or firms are denominated as “bridges” because they link two important groups. Bridges may not be very central according to all the other measures but have high levels of betweenness. All of our measures of centrality are normalized by the number of total connections existent in the network, which generates an index where the total number of possible connections equals 100, so a figure that has a degree of 55 has 55% of all the possible connections it could get.

Finally we use a more sophisticated measure of centrality called eigenvalue. That value is a recursive measure of centrality. This indicator takes into account the connections of a point X in the first place, then it takes into account how connected were all the connections of that point X, that is how connected were the points Y, W, Z, K that are connected to X. Then it recursively sees how connected were the points connected to Y, W, Z, and K. This yields a more robust measure of centrality. This indicator is also normalized as an index. For a more detailed description of centrality measures see Scott (1991).

The problem with all these measures of centrality is that they only work for a closed network. Networks fragmented into clusters cannot be studied as one for the purposes of studying centrality. Then it was hard to compare the Mexican and Brazilian networks in that respect.

[10] Our data on federal delegates is comprehensive but at the state level data was more difficult to obtain. We would expect that poltical participation to increase slightly with this information. Such data, however, would still confirm that elite networks were more clustered in Brazil than in Mexico.

[11] Many of these arguments center on a group called los científicos (the scientists) who masterminded much of Porfirio Diaz’s policies and helped him retain power.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download