Elite studies, even in an age of “little people’s” history, require no apology to the degree that they reveal patterns of social stratification and the dynamics of power, both economic and political. The elite examined here is the political leadership of São Paulo state between the coup d’état establishing the Brazilian Republic and that of the centralizing Estado Novo (1889-1937). This study reveals the close correspondence between political and economic leadership, present to a lesser degree in analogously defined elites of Minas Gerais and Pernambuco, but apparently lacking in other periods of Brazilian history and in elites of other countries. It also asks why such a correspondence exists, and looks at the character of the dominant social class, as well as the question of hegemony, in the light of recent modes-of-production debates about São Paulo and other regions of Latin America.1

For the years under study, Brazil had 20 states, but São Paulo played a crucial role in framing the terms of Brazilian politics. The Republican party of São Paulo (Partido Republicano Paulista) was the largest and best organized at the fall of the empire (1889), and a Paulista presided over the constituent assembly of 1890–91. São Paulo provided the first three civilian presidents between 1894 and 1902. It furnished more chief executives (4) until the fall of the Old Republic in 1930 than any other state machine, and two others who were elected but did not serve.2

São Paulo’s political leadership was related to its economic preeminence. From 1889 to 1937, coffee produced 40 to 70 percent of Brazil’s export earnings, and São Paulo accounted for the bulk of the country’s coffee; in fact, the state produced half the world’s total in the decade after 1910. São Paulo also acquired its industrial leadership in the period studied, accounting for one-third of Brazil’s industrial production in 1919 and 43 percent by 1939, the year of the next economic census. Moreover, the fiscal responsibilities of the state were unique: both its outlays and its foreign debt were 25 percent as large as those of the federal government by the depression years. Finally, São Paulo is notable for having absorbed about half of Brazil’s immigrants in the era under consideration, the country’s only period of mass immigration; this fact, plus internal migration from other states in the 1930s, helped it become the most populous state in the country at the end of the period.3

This study builds on (and adds to) data sets and interpretations by John D. Wirth, Robert M. Levine, and the two authors of this article.4 In each of these studies, a political elite was defined with respect to a set of formal positions judged to be relevant to political power and patronage in Brazil, at state and federal levels in the years 1889-1937. Thus, like virtually all other historical elite studies, this one employs a positional rather than a reputational or decision-making means of defining a political elite.5 Specifically, it regards as members of São Paulo’s political elite the holders of 18 important positions in the state government and state-based parties—no national party in this period survived a presidential election—and Paulista office holders in 17 important federal posts, from the birth of the republic to the Estado Novo. In all, 263 individuals were identified as forming the elite in question.6

In a related paper, two salient characteristics of the São Paulo elite emerged. The first, the relatively large number of property owners present, will be treated in detail below. The second, also shared by analogous elites in Minas Gerais and Pernambuco, is the high proportion of holders of university degrees—93 percent in the case of São Paulo. When compared to political elites in Mexico (1900-71), Argentina (sample years of 1889, 1916, and 1946), and those of the United States and Europe in the nineteenth and twentieth centuries, the three Brazilian state elites top the list in the share who attained university education. This phenomenon has a counterpart in other periods of Brazilian history: highly educated political elites led the empire (1822-89) and predominated in the period after World War II as well.7

Given this pattern among the Paulistas in an age when few could afford higher education—a member of the Paulista political elite was 243 times more likely to have had a university education than the state population as a whole as late as 19408—it is not surprising that data on father’s occupation, the single best proxy for social class on entry into an elite, revealed that 95 percent of the São Paulo group came from upper-middle or upper-class families. This figure is high compared to modern European elites; even in Porfìrian Mexico (data for 1900-11), Peter H. Smith found that 11 percent of the political elite had fathers with lower-class occupations (worker, peasant).9

It is instructive to compare the occupations of the São Paulo elite (as opposed to those of the fathers of elite members) with occupations of the elites of Minas and Pernambuco, for which identical definitions and coding conventions were used. The elite of São Paulo led those of the two other states in having the largest shares of fazendeiros (owners of market-oriented estates), industrialists, merchants, bankers, rural land dealers, Comissários (short-term lenders to fazendeiros), and investors in railroads. Of the occupations associated with property ownership, Minas and Pernambuco outpaced São Paulo only in the category of mine owners (São Paulo had none). The Paulistas led the other two groups in the composite variables BUSINESSMAN (41 percent of the Paulista elite), and PROPRIETOR (56 percent); São Paulo also led in the share of its elite associated with the agricultural export complex (40 percent).10

Comparison of the Paulista elite’s occupations with those of elites in other countries and in Brazil in other periods is complicated by the fact that some students of political elites have used multiple codings of occupations, while others have not. Furthermore, definitions are not identical for such terms as “farmer,” “landowner,” and “businessman.” Finally, São Paulo was a constituent state in a federation, whereas the nation-state is the unit of analysis in the other studies considered. For these reasons, comparison can only be suggestive. It must also be emphasized that we have not attempted to compare economic elites as such. Definitions of these elites are probably more subjective and therefore more likely to vary from study to study than those of political elites, at least the ones defined in positional terms. Rather, the question is the degree to which political elites were penetrated by property owners.

Professionals seem to predominate everywhere, but São Paulo’s elite has the highest share of “economic” or property-owning occupations among the cases examined; as noted, 56 percent of the Paulista elite had occupations in which income took the form of profits, interest, or rent, rather than salaries or fees for services.11 This is the sense in which “proprietor” and “owner” are used below. The Paulistas contrast sharply with the Brazilian imperial elite: José Murilo de Carvalho’s ministerial elite contained only 5 percent proprietors.12 Similarly, in Smith’s Mexican political elite before, during, and after the Mexican Revolution, landowners never rise to 4 percent of the total elite, and businessmen only reach 6 percent, as contrasted with 38 percent and 41 percent, respectively, in the São Paulo elite.13 For the last parliament of the July Monarchy in France (1846-48), Patrick and Trevor Fligonnet found that only 13 percent of Guizot’s government deputies were businessmen. Of all bourgeois deputies, only one-tenth were in business occupations. They comment: “This last figure is surprisingly small for a regime that was . . . described by Marx and Tocqueville as a joint stock company ruling the many for the material advantage of a selected few. . ..”14

Two other parliamentary elites with similar characteristics are the Spanish and Dutch chambers of deputies. In seven Spanish parliaments, from 1879 to 1979, “businessmen” and “farmers,” the only two property-holding groups identified in a study of this elite, together averaged 17.7 percent of the total membership. In the Dutch chamber for the years 1849-1967, the average share of businessmen and farmers was approximately 13 percent.15

Most curious of all, perhaps, is the case of the U.S. political elite (1877-1934), defined as presidents, vice-presidents, and cabinet members, for which Robert D. Putnam found only 13 percent were businessmen and 2 percent were landowners. For the longer-term U.S. national leadership (from the 1790s to 1940), Putnam remarks: “The proportion of businessmen (or sons of businessmen) entering the political elite has remained relatively small and essentially unchanged throughout the period during which America moved from an overwhelmingly agricultural society to an overwhelmingly industrial society.”16 These data are in apparent conflict with an earlier study of U.S. cabinet members, in which a majority were identified as businessmen for the period 1889-1949. Yet, the latter study also shows that professionals rather than property owners dominated the executive elites (national cabinets) of Britain, France, and Germany.17

If the conflicting findings for the more narrowly defined U.S. elite (i.e., an exclusively executive elite) are excepted, the low level of participation of proprietors is notable everywhere outside Brazil.18 The only group treated above that approaches São Paulo’s is Argentina’s; Dario Canton’s proprietors among the members of the three Argentine parliaments range from 24 to 45 percent, and a weighted average is 31 percent. This figure is still considerably below the São Paulo score. In José Luis de Imaz’s Argentine elite study, Los que mandan, his “governing teams” include “entrepreneurs,” the only group of proprietors he identifies, which presumably contains landowners. The entrepreneurs’ share of the total (in a multicoded occupational scheme) ranges as follows: 16 percent in 1936; 8 percent in 1941; 10 percent in 1946; o percent in 1951; 24 percent in 1956; and 32 percent in 1961.19

Thus the São Paulo elite seems to have a large proportion of direct ties with the economic sector. Given the second of the two U.S. studies cited, it may still be objected that if property owners are more heavily represented in executive elites, the mixed set (executive, legislative, and judicial) for São Paulo is too heavily weighted in that direction. Yet the strictly parliamentary subset had an even higher share of property owners than the whole; i.e., proprietors formed 61 percent of the São Paulo legislative membership, as opposed to 56 percent of all Paulistas in the political elite.20

If the Paulista political elite for 1889-1937 was apparently unique in the coincidence of property owners and political leaders, it would be nonetheless incorrect to infer that it was unique in its close relations with propertied groups. Though property owners tend not to predominate in Western political elites, extensive research has nonetheless confirmed the unrepresentative nature of such elites. Numerous studies have shown that political elites tend to be made up of professionals from relatively privileged backgrounds, linked with owners of property through social and economic ties, or (in much smaller numbers) are made up of the owners themselves.21 As Ralph Miliband, probably the most influential Marxist student of elites, puts it, “The reason for attaching considerable importance to the social composition of the state elite in advanced capitalist countries lies in the strong presumption [of class partisanship] this creates in its general outlook, ideological dispositions, and political bias.”22 This observation seems a fortiori valid for the Brazilian elite under study, which manifestly felt no compulsion to represent the interest of the laboring classes of city and countryside, not to mention other groups it all but excluded—nonwhites, women, and immigrants. (Among the 263 São Paulo elite members, there were two nonwhites, one woman, and one immigrant.)

The indicators of social class alluded to above—occupation and father’s occupation—reveal the Paulista political elite as a relatively impenetrable group firmly entrenched in economic activities. The question thus arises whether São Paulo had a “power elite” in the sense in which C. Wright Mills used the term: were there overlapping political and economic elites tending to monopolize both types of power at the highest levels?23 One need not expect, however, that the capitalist class as such would directly take charge of government. Early in the century, both Karl Kautsky and Max Weber made the point that modern business enterprise is too time-consuming an activity to permit capitalists to undertake the complex tasks of government. Miliband believes that relatively “pure” cases of rule by the capitalist class are found only in the early modern period, in such city-states as Lübeck.24 According to Mills, what we might expect to find, instead, is a web of cliques with the same class background, sharing the same churches, schools, or social clubs, or the same types of such institutions.

Among the studies listed in note 7, only Smith formally treats the issue of a power elite. For the years he studied, 1946-71, Smith found virtually no overlap of political and economic elites. In addition, for Brazil in the 1970s, McDonough implicitly denies the existence of a power elite.25 (If true, the absence of a power elite does not preclude the intermittent participation in politics by capitalists as such, as René Dreifuss has recently documented with respect to Brazilian industrialists in the crisis of populism and its resolution in the coup of 1964.)26

Although data on São Paulo’s economic elite for the years under study can only be suggestive, three partial but independent forms of evidence permit a comparison of the economic and political leadership. The first, which can be treated briefly, consists of data in a new study examining the holdings of the richest men whose estates were inventoried in the Primeiro Ofício da Família de São Paulo between 1870 and 1895. Of the ten wealthiest, seven were relatives of the state’s political elite as defined here, and another was a business partner of an elite member.27 Thus four-fifths of the ten wealthiest Paulistas sampled at the outset of the period had intimate connections with the political elite; the fact that none of the ten held an elite-defining office is at least partly explicable by the fact that four had already died before the republic was established in 1889, and the others were aged.

The second body of evidence permits a more formal test for a power elite in Sao Paulo. The evidence is provided by a directory of prominent businessmen and professionals in the state in 1932 compiled by the British Chamber of Commerce of São Paulo and Southern Brazil.28 The test asks whether the political leaders are also business leaders, and in addition, whether the former’s business associates and relatives appear in the list of business leaders. (This is a stronger test than the one proposed by Mills and used by Smith, viz., to ascertain whether persons in the business and political elites have the same sorts of backgrounds, such as having fathers in the same occupations, attendance at the same or similar schools, and membership in the same churches and social clubs.)29

Not all of the men listed in the directory, Men of Affairs, are identified as property owners (though some not so identified in fact were). Therefore, we have specified three groups of political elite members, corresponding to stricter and looser associations with the economic elite.

  1. A core group consisting of political elite members who are identified as property owners in Men of Affairs;

  2. Political elite members whose identifiable property-owning relatives appear in the book, as well as those whose business partners are listed;

  3. A group of professionals (mostly lawyers) consisting of political elite members who were listed in Men of Affairs without reference to property, and of those whose identifiable relatives appear in the directory without mention of ownership. We assume that these professionals served the interests of property owners and that they can therefore be considered part of a broad economic elite.30

These three groups were compared to the relevant pool of political elite members (of a total population of 263), i.e., the 169 Paulistas who were alive in 1932, the year the data for Men of Affairs were gathered. The results are as follows:

Obviously, to set a given level of coincidence of the two elites as proof of a power elite in Mills’s sense would be arbitrary, but an overlap of 60 percent appears significant, especially since Mills himself found a 57 percent overlap in his own study.31 If we further recall that these groups are more strictly defined as overlapping political and economic elites than those of Smith, it seems highly probable that a power elite in Mills’s sense existed for São Paulo during the era in question.

An additional inquiry concerns the degree to which these 102 members of São Paulo’s economic elite succeeded within the political elite. If business leaders were more important in political leadership than nonbusiness leaders, then the economic elite groups above should have “succeeded” more fully than other members of the political elite. The greatest success, furthermore, should be found among members of the propertied groups (groups 1 and 2). Following a procedure used in Smith’s Labyrinths of Power, the elite-defining positions were stratified to ask the question: Is success correlated with property ownership? For this purpose, the posts were arrayed from lowest (1) to highest (6), using the following categories:

(1) state legislative and executive posts other than department heads (i.e., secretaries); (2) state secretaries and presidents of the state courts; (3) executive committees of the major parties in São Paulo— the Partido Republicano Paulista, the Partido Democrático, and its successor, the Partido Constitucionalista; (4) federal legislative and executive posts other than ministers; (5) governors and supreme court justices; (6) presidents, vice-presidents, and ministers.32

This ordinal variable, HIGHEST OFFICE, can be correlated with property-owning occupations of elite members. “Success,” however, need not be measured in terms of the highest office attained, and an alternate definition would be the number of elite-defining positions held. Thus, another measure of success can be used, namely, NUMBER OF OFFICES.

To measure the degree of success on the two scales, gamma was used. This is one of the several possible tests of strength and direction of correlation (as opposed to significance). Varying from +1 to −1, gamma coefficients can test the statistical relationship between economic leadership and political leadership.33 If a power elite existed, then membership in the “economic” groups should be positively correlated with success. The results of such tests are summarized in Table II.

These results in general correspond to what we should expect if a power elite existed in São Paulo: membership in the overlapping political and economic groups tended to be positively associated with success. Moreover, the relationship was stronger when professionals were excluded from the overlapping elite, and the core group of property-owning elite members was the most closely associated with both measures of success. Overall, there seems to be a stronger correlation of business connections with survival in the elite (number of offices) than with movement up the ladder.

The relationship between political success and business leadership can perhaps be seen even better by noting the mean scores on the two scales of success for each of the groups:

The relationship is clear: with a closer connection to the state’s economic elite, a political elite member could, on average, expect to reach higher offices and to hold a greater number of posts.35 On both scales of success, the central property-owning group had mean scores fully half a step higher or farther than politicians not included in the overlapping political-economic elite.

A further and independent test of the association between business leaders and political leaders involves the 9 percent of our entire São Paulo political elite who had known foreign business ties. These 23 persons were all investors in, or managers of, foreign firms or branch operations in Brazil. It is reasonable to assume that these foreign-connected business leaders were in the highest stratum of the Brazilian bourgeoisie.36

For these men, the gamma coefficients were +.41 for NUMBER OF OFFICES and +.41 for HIGHEST OFFICE. Their mean (absolute) scores on NUMBER OF OFFICES was 3.04, compared to 1.94 for all the rest of the São Paulo set; it was 3.65 for the 23 associated with foreign firms on the HIGHEST OFFICE variable, compared to 2.86 for all other members of the São Paulo political elite.

Three different forms of evidence thus support the assertion of a strong relationship between ownership of income-producing property and political leadership in São Paulo. The first two show a close association between property owners and the political elite, while the second and third, using two definitions of “success,” demonstrate statistically that property ownership was related to the degree of success, once politicians had entered the elite. Under such circumstances, it does not seem farfetched to call the Paulista political elite an “executive committee of the ruling class.”

Of course, these findings would not be especially important if vested interests—especially those of the coffee complex—had not benefited disproportionately from government policies at state and federal levels. But they did so benefit, as will be argued below. Nor do the findings imply that education was unimportant. Jorge Tibiriçá’s political success—he was the father of coffee valorization and twice governor of São Paulo—may have depended in some degree on his educational attainments: two doctorates from European universities made him the most “educated” person among the 263 members of our Paulista political elite. It is also relevant that there was a perfect negative correlation between the lack of a university education and the number of offices held; i.e., all those who lacked a university degree only held one office. (Perhaps it was a lack of university education that kept the immigrant bourgeoisie out of political office, rather than their full-time devotion to industrial management.)37 In any case, education was to some degree dependent on wealth, and it was the combination of ownership and higher education that distinguished the São Paulo political elite in this period. (Tibiriçá, once again, was immensely wealthy.)

Despite the importance of education, economic prominence clearly appears to be a more distinctive feature of the Paulistas, at least as compared to other Brazilian political elites. Furthermore, except for the conflicting findings on U.S. elites, the tendency towards a coincidence of economic and political leaders was greater than among the other modern political and economic elites of which we have knowledge, including Brazilian elites in earlier and later periods.

To many readers, the explanation of such a correspondence is perhaps a more interesting problem than establishing its existence. Following Miliband, one might, for example, seek an explanation in the “nonmodernity” of the economic and political systems.38 It is clear from the profile of the political elite of São Paulo that its constituents lacked the specialized functions characteristic of modern industrial society, since each person had an average of three occupations, in addition to that of “politician.” Then, too, members of the São Paulo political elite attained power under a regime with sharply restricted suffrage in a period when the Paulista economy was undergoing its most rapid advances, responding to the impetus of the “second industrial revolution” in the North Atlantic basin. An effective suffrage of 2 to 7 percent of the Paulista population prevailed at a time of rapid economic growth, perhaps contributing to the close correspondence of economic and political elites. But such facts do not offer a sufficient explanation, for Cantón and Smith found considerably smaller numbers of property owners in Argentina and Mexico in their respective elites of 1889 and 1900-11;39 their findings were similar to those on other countries. For Cantón’s group, in fact, the share of proprietors even rose in the first parliament elected on the basis of obligatory male suffrage. Thus the explanation appears not to lie exclusively in the combination of rapid economic growth and low political mobilization, although Smith suspects there was a power elite in Porfirian Mexico.40

An additional explanatory factor for São Paulo might be the economic opportunities associated with the institution of a federal regime. One of these was a radical shift of revenues to São Paulo from the central government. Between 1891 and 1892, the first year after the ratification of the new federal constitution, São Paulo’s revenues rose from 4 percent to 17 percent of the level of federal revenues, and expenditures followed a similar path, from 3 percent to 12 percent; by the years 1930-37, revenues had reached 23 percent of those of the federal government, and expenditures, 25 percent. Meanwhile, real central government revenues and expenditures fell in per capita terms in the early years of the new system.41 In addition, the republic permitted São Paulo to acquire a large foreign debt, which twice in the period studied reached one-third the size of the federal debt.42

The opportunities these fiscal shifts presented may have attracted economic actors like Jorge Tibiriçá, Rodolfo Miranda, João Rubião Junior, and other wealthy fazendeiros and hankers who successfully sought to structure governmental authority for the interests of their class. A related consideration is the devolution of the public domain from the central to the state governments under the republican constitution. Political connections were extremely important in gaining control of land in frontier areas, which included most of São Paulo at the outset of the republic.43

An examination of the nature of the Paulista political elite should also ask whether São Paulo’s economic elite, which penetrated the political elite so strongly, was a capitalist class, i.e., a modern bourgeoisie, or something else. If the Brazilian economic elite in our period were precapitalist, then a model of feudalism might be appropriate—one in which the class that controlled the land was also the ruling elite (as Poulantzas emphasizes).44 There is in fact an analogy between the lack of absolute property ownership in the feudal period of European history and the inability of the Brazilian state in the nineteenth and twentieth centuries (São Paulo included) to assure title to rural property.45 But the feudal scheme breaks down when one recalls that under the empire, as noted, the ruling elite apparently consisted of far fewer property owners than the republican elite (though Carvalho shows a tendency toward greater association of political leadership with ownership as the empire grew older).46 More importantly, the rural labor system, at least in São Paulo (see below), had features lacking any equivalent in Marcello Carmagnani’s well-known model of Latin American feudalism.47

It is also known that the Paulistas exhibited traits of modern entrepreneurship in the sense of buying and selling real property, a fact related to their frequent change of residence.48 In addition, on the whole, Paulista political leaders were far from being a long-established hereditary group (quatrocentão): the single most frequent surname among the Paulista elite (borne by six members) was Rodrigues Alves, whose common forebear arrived on Brazilian shores in 1832.49

If one considers the general economic context, moreover, and not just the dominant class, basic features of capitalism appear to prevail. After the end of slavery in 1888—and more universally after new labor laws in 1890—not only was plantation labor employed on the basis of formal contracts, but “contract enforcement and labor discipline were . . . to depend on the mechanism of the market,” in the words of Verena Stolcke and Michael M. Hall.50 Thomas H. Holloway has shown that the “unique colono system” the planters created, with the economic feature of usufruct as well as wages, generated a dynamic labor market, often viewed as a characteristic feature of capitalism.51 Strikes, a feature of capitalist labor relations, were surprisingly frequent in rural São Paulo; several dozen occurred between 1890 and 1913, and over a hundred followed before the end of the Old Republic.52

Paulista political and economic leaders pursued a policy of maximizing the number of available workers, and workers responded by moving from estate to estate. Geographical mobility was probably at least as common among the immigrant coffee workers as among the elite.53 At the same time, the dense colono population and its prosperity developed a domestic market that made possible the rise of an industrial society. These features, too, would seem to place the Paulista case in the realm of modern capitalism. Granted, Paulista planters tried to “fix” immigrants to the land,54 but it didn’t work; the only effective control over labor resulted from state-supported—nay, state-induced!—saturation of the labor market through the subsidization of immigration.55

The question remains whether any estate-owning class and its political representatives can really be considered “capitalist” or “bourgeois” in the accepted sense. Yet E. P. Thompson believes, like Karl Marx, whom he cites on this point, that there can be no doubt about the agrarian capitalists of early modern England.56 Clearly, the planters of São Paulo in the period under review were in a different political situation, facing no major domestic rivals; but the Paulistas would not seem less bourgeois for emulating triumphant agrarian bourgeoisies abroad, as indicated by the formation of the Sociedade Rural Brasileira (and its predecessors), modeled on Britain’s Rural Society.

In Brazil, among those who see the bourgeoisie as dominant, much recent debate has revolved around the issue of hegemony: was the state relatively autonomous from the dominant class?57 For the case of São Paulo in the years 1889-1937, we suspect that the answer is “no,” based on the degree of coincidence of ownership and leadership, as well as on specific government policies.58 This view is at odds with those found in new studies by Mauricio Font and Steven Topik, both of whom address the hegemony issue, though from different perspectives. Font argues that São Paulo’s planters were losing their hegemonic position within São Paulo in the 1920s, while Topik denies that coffee interests were “hegemonic” at the federal level in the Old Republic.59

Focusing on the period 1920-30, Font argues that São Paulo’s planter elite, which he views as having exercised a kind of Gramscian hegemony in earlier decades of the republic, was losing control of government at state and national levels. In part because of dissension among large planters, power was passing to industrialists, small farmers (including coffee growers), and other economic groups, as well as to an increasingly autonomous government led by the Partido Republicano Paulista (PRP) at state and national levels.60 Font thus extends back into the 1920s the thesis of Francisco Weffort and Boris Fausto about the Estado de Compromisso of the Vargas years—in which no group had absolute authority comparable to that enjoyed by the planters (in the Weffort-Fausto view) before the fall of the Old Republic. From this it follows that the state had a significant measure of autonomy from economic interests.61

Font holds that, over the course of the 1920s, the bases of planter rule were undermined by the lack of a saturated rural labor market, by the long-term decline of the comissário credit system, and by the decline of clientelism and coronelismo (rural bossism) as forms of interest aggregation. He further argues that an “alternative economy” appeared in the 1920s, based on the rise of small farms (frequently owned by immigrants), and on a rising share of economic activity by domestic commerce and industry. Industrialization, in fact, is seen as challenging, rather than dependent on, coffee capital. Font interprets the relationship between the planter elite and the state and federal governments as one of conflict, revolving around economic policies which did not favor large planters.62

Font’s data and reasoning are convincing with respect to the existence of significant disputes among large planters and the growing importance of nonmagnate planters in the coffee economy.63 In addition, an analysis of our political elite data by decade shows a somewhat smaller share of fazendeiros in the 1920s than in earlier decades, as Font’s thesis would indicate.64 This fact is not surprising, given the increasing differentiation of the economy (and specifically the shift toward industry as the century advanced).

The argument is not wholly convincing, however. First, Font does not take into account the lack of unity among planters even in the period of their alleged hegemony, since his detailed study is confined to the decade of the ’20s.65 Second, many of the planters and their associates, whom he notes as contesting state coffee policies in the 1920s, were dissenters from the PRP in previous intraparty contests, going back to the “schism” of 1901.66 Third, in treating dissatisfaction with the PRP, he fails to take into account the generational cleavage that had developed by the 1920s.67 Finally, the thesis of planter conflict with state authority runs into the problem of government “output.” São Paulo’s coffee debt, which was virtually eliminated in 1921, was renewed in 1926 (with the establishment of the state bank) and had grown almost as large as the noncoffee state debt in 1930; in addition, the total coffee debt tripled from 1929 to 1930. Almost one-half of São Paulo’s foreign debt in 1930—one-third as large as that of the national government—was contracted directly for the support of coffee valorization.68

Consequently, the close fit between economic and political elites, together with concrete government policies, seems to support the view that economically powerful groups, led by those associated with coffee interests, controlled the state government.69 To the degree—if any—that control of coffee policy was passing from the planters to a semiautonomous state government, such a process probably had more to do with demands from foreign lenders for government controls over coffee production and storage than with challenges within the São Paulo economy and polity.70

Addressing the problem of planter influence at the federal level, Topik shows important constraints on the power of coffee and other export interests. Yet one likely reason for the gap that he notes between the expectations of exporting groups and government action on their behalf is the mere fact that Brazil’s was a “soft” state, in Gunnar Myrdal’s phrase, with limited capacities to implement policies of almost any kind.71 Furthermore, the power of the coffee complex in the “state” looks quite different if the state is defined to include the São Paulo government, which, for example, acquired a coffee valorization debt in 1908 for 15,000,000 pounds sterling, a sum half again as large as the federal government’s onerous Funding Loan ten years earlier.

Topik notes that obligations by coffee groups to the federal government were repaid during the Old Republic, but he does not consider the state government’s huge foreign loans, which were not. Here Topik’s conclusions are strongly influenced by his periodization, ending in 1930. In the succeeding three years, the federal government assumed joint responsibility for the coffee debt São Paulo acquired in the 1920s, and favored the planters by introducing the Reajustamento Econômico, canceling half their mortgage debts and limiting mortgage interest rates. The Bank of Brazil clearly favored São Paulo over the other states both in loans to the state government and in private loans for the years for which data are available, 1933-37. By October 1937, the outstanding Bank of Brazil loans to the federal government for coffee defense amounted to more than half the federal total for all purposes.72 For these several reasons, the implied “relative autonomy” of the state, as advanced by Topik, seems questionable.73

In conclusion, several characteristics of the São Paulo political elite stand out. It was relatively impenetrable to women, nonwhites, and immigrants, and was narrowly recruited among a small pool of highly educated men, who, more often than not, were property owners. Although Brazilian political elites in other periods and regions exhibit high rates of formal education compared to non-Brazilian elites, their relative shares of proprietors were apparently smaller than the São Paulo elite’s. The importance of ownership among the Paulistas also distinguished them from political elites in other countries in modern times, and this coincidence between economic and political leadership seemingly points to the existence of a Millsian “power elite.”

Among the several plausible explanations of the coincidence of elites, the possibility of a premodern economy is hard to reconcile with the existing relations of production in the coffee sector. A restricted suffrage, combined with rapid economic growth, may have been a contributing factor. But the change of regime from empire to republic must also be taken into account. The partial devolution of fiscal authority to the states and the devolution of the power to validate land titles may have offered unique incentives for economic actors in São Paulo to play political roles. The coffee frontier was in its heyday, and the state government offered a means of securing foreign financing for coffee valorization, which the risk-averse federal government was long reluctant to undertake.

Whatever may be the precise reasons for the observed pattern, it does call into question the “relative autonomy” of the state. That appears to be a dubious thesis for the government of São Paulo in the 1920s, and it is a far from obvious one at the federal level down to 1937. In effect, the Paulista economic and political elites made repeated claims on government intervention of various sorts, such as valorization, exchange rate manipulation, and loans from the Bank of Brazil. Such demands were successfully pursued by an integrated, highly educated ruling and owning group that shaped the state (at all levels) to its own regional needs, a fact which may in part account for the economic leadership, phasing into dominance, that São Paulo established in the period studied and still persists today.


For Brazil, see José Roberto do Amaral Lapa, ed., Modos de produção e realidade brasileira (Petrópolis, 1980). On pp. 29-32 Lapa provides a list of 60 works, organized by type of interpretation—feudal, capitalist, or uniquely Brazilian modes of production; a fourth category of studies revises the others. Among U.S. Brazilianists, Barbara Weinstein has entered the modes of production debate with The Amazon Rubber Boom:1850-1920 (Stanford, 1983). José Carlos Chiaramonte has recently reviewed the literature on modes of production in Spanish America (treating Sergio Bagú, André Gunder Frank, Rodolfo Puiggrós, and later writers) in Formas de sociedad y economía en Hispanoamérica (Mexico City, 1984).

Of special interest in this study are works by Mauricio Font and Steven Topik, which address a specific aspect of the modes of production problem in Brazil—the relation between the export-oriented bourgeoisie and the state. See Font, “Padrões de ação coletiva dos plantadores paulistas de café: 1932–1933,” in Economía e movimentos sociais na América Latina, Fernando Henrique Cardoso, Bernardo Sorj, and Mauricio Font, eds. (São Paulo, 1985), pp. 218-244; and Topik, “The State’s Contribution to the Development of Brazil’s Internal Economy, 1850-1930,” HAHR, 65:2 (May 1985), 203-228.


Francisco Rodrigues Alves, president between 1902 and 1906, was reelected in 1918, but was too ill to serve and soon died. Júlio Prestes was elected in 1930, but the Old Republic (1889-1930) was overthrown a month before he was to take office. For São Paulo’s share of other key executive, judicial, and congressional posts, see Joseph L. Love, São Paulo in the Brazilian Federation, 1889-1937 (Stanford, 1980), pp. 185-188.


Ibid., pp. 1-68, 240-266.


The studies from which our data are taken are Robert M. Levine, Pernambuco in the Brazilian Federation, 1889-1937 (Stanford, 1978); John D. Wirth, Minas Gerais in the Brazilian Federation, 1889-1937 (Stanford, 1977); Love, São Paulo; and Love and Bert J. Barickman, “Regional Political Elites in Brazil: A Comparison of Three States, 1889-1793,” forthcoming.

We are grateful to Professors Levine and Wirth for supplying additional data on father’s occupation for this study, and for their comments on a draft of this paper.


On the three basic methods of defining elites, see Robert A. Dahl’s classic article, “A Critique of the Ruling Elite Model,” The American Political Science Review, 52:2 (June 1958), 463-469.

Defining an elite on the basis of reputation of power wielders, by an informed panel of judges, has occasionally been done; Frey and Bonilla, for example, use this method in combination with the positional means. The technique is ofless value to the historian than to the political scientist, who can rely on contemporary judges. The decision-making means of defining an elite would be useful to the historian only for one issue at a time, and for many issues records do not exist or are inaccessible. See Frederick W. Frey, The Turkish Political Elite (Cambridge, MA, 1965); and Frank Bonilla, The Failure of Elites (Cambridge, MA, 1970). A contemporary community study in Minas Gerais employing the three methods is that of Celson José da Silva, Marchas e contramarchas do poder local (Belo Horizonte, 1973).


In the state, the list consisted of the governor and his important elected and appointed assistants in the executive branch, together with top-ranking legislative and judicial figures. Federal posts were defined analogously, but were included only when held by Paulistas. Since some persons held office for as little as one day, a 90-day minimum tenure was required for inclusion during the period Nov. 15, 1889 to Nov. 10, 1937. Specifically, state elite members consisted of the governor; the lieutenant governor; the secretaries of justice, finance, agriculture, transportation, education and health, security, and the governorship; the state chief of police; the president of the state bank; the prefect of the state capital; the president of the Coffee Institute of São Paulo; the presidents of the state chamber and senate; the majority leader in the chamber; and the president of the state supreme court.

Federal officials included the president of the republic; the vice-president, the ministers of justice, finance, agriculture, transportation, education, labor, and foreign affairs; the president of the Bank of Brazil; the prefect of the Federal District; the president of the National Coffee Department; the president of the federal chamber of deputies; the vice-president of the senate (the presidency of that body being exercised by the vice-president of the republic); the majority leader of the chamber; the leader of the state delegation in the chamber; and members of the federal supreme court.

The elites of Pernambuco and Minas Gerais were defined in a like manner. In fact, the elites were defined uniformly except for “leading (state) administrators.” To have eliminated such key positions as the presidencies of the Coffee Institutes in São Paulo and Minas because the post did not exist in Pernambuco seemed unduly procrustean, since the central purpose of the definition of elites was to obtain comparable wholes from each state. The size of the elite populations ranged from 214 in Minas Gerais, to 263 in São Paulo, to 276 in Pernambuco.

Three nongovernmental positions are included in the three elites—membership on the executive committee of the Republican party (the “establishment” party in the Old Republic), and on those of the most important non-Republican party before 1930, and from 1931 to 1937. (The roles of the Republicans and their opponents were reversed after 1930, or new establishment parties brought in many leaders of pre-1930 opposition groups.) It seemed essential to include this nongovernmental sector of the elite because of its leading role in the larger political process. Separate computer runs of party executive committee members against nonmembers tended to confirm our qualitative judgment on this point: committee members held more elite-defining positions than nonmembers, and they were more prominent in the economy and society. Thus, in the manner of Bonilla’s definition of the Venezuelan political elite, in The Failure of Elites (p. 16), party executive committees were included to better approximate the effective power structure. Regarding occupations, if any biographical source indicated that an individual had more than one, all were coded. For further discussion of classification procedures and definitions of occupations and other background variables, see Love, São Paulo, pp. 277–287.


For details and further interpretation, see Love and Barickman, “Regional Political Elites. ” See also Peter H. Smith, Labyrinths of Power: Political Recruitment in Twentieth-Century Mexico (Princeton, 1979), p. 85; Darío Cantón, El parlamento argentino en épocas de cambio: 1890, 1916 y 1946 (Buenos Aires, 1966), p. 38; Frey, The Turkish Political Elite, p. 44; Robert D. Putnam, The Comparative Study of Political Elites (Englewood Cliffs, NJ, 1976), p. 27; Wolfgang Zapf, Wandlungen der deutschen Elite: Ein Zirkulationsmodell deutscher Führungsgruppen, 1919-1961 (Munich, 1965), p. 178; Eul-Soo Pang and Ron L. Seckinger, “The Mandarins of Imperial Brazil,” Comparative Studies in Society and History, 14:2 (Mar. 1972), 235; José Murilo de Carvalho, A construção da ordern (Rio de Janeiro, 1980), p. 63; Peter McDonough, Power and Ideology in Brazil (Princeton, 1981), p. 62; and David V. Fleischer, “Thirty Years of Legislative Recruitment in Brazil: An Analysis of the Social Backgrounds and Career Advancement Patterns ot 1548 Federal Deputies, 1945-1975” (unpublished paper presented at the Tenth World Congress of the International Political Science Association, Edinburgh, Scotland, 1976), p. 29.


The national figure was one graduate for every 387 Brazilians. Data from Instituto Brasileiro de Geografia e Estatística, Recenseamento geral do Brasil (1° de setembro de 1940), série nacional, vol. 2: Censo demográfico: População e habitação (Rio de Janeiro, 1950), pp. 1, 30 67, 89.


For São Paulo, data on fathers’ occupations were obtained for 130 elite members. Details are presented in Love and Barickman, “Regional Political Elites.” Unfortunately, few elite studies provide fathers occupation. One which does is Zapf’s on Germany. In three regimes (Weimar, Nazi, and Federal Republic), fathers of political elite members had lower-middle and lower-class occupations in the following shares: 47 percent (1925), 59 percent (1940), and 54 percent (1955). Lower-class occupations alone were 14, 5, and 16 percent, respectively. In Mexico, the share with lower-class fathers rose to 17 percent in the revolutionary elite (1917-40) and to 24 percent in the postrevolutionary elite (1946-71). See Zapf, Wandlungen der deutschen Elite, pp. 180-182; and Smith, Labyrinths of Power, p. 77.


Scores for Pernambuco and Minas were 24 and 35 percent on BUSINESSMAN, respectively, and 37 and 45 percent on PROPRIETOR. BUSINESSMAN is a composite variable, consisting of those who were merchants, industrialists, bankers, Comissários, mine owners, or investors in railroads. Members of the agricultural export complex were those who were agricultural society officers, exporters, fazendeiros, or Comissários. PROPRIETORS consisted of those who were fazendeiros, Comissários, merchants, industrialists, bankers, rural land dealers, railroad investors, mine owners, importers, exporters, and investors in foreign firms operating in Brazil. For explicit definitions of occupations, see Love, São Paulo, pp. 283, 285.


Some property owners could have been “front-men” (testas de ferro) for others; only a study of hundreds of estate inventories (many of them not yet available) could definitively resolve the matter. In any event the principal cause of the front-man phenomenon in many Latin American countries today—restrictions on foreign ownership—was lacking in the period studied.


Carvalho, Construção da ordem, p. 79. Using a broader set of criteria on p. 87, Carvalho presents data on ministers who were “linked” with property as owners, or as sons or sons-in-law of owners; 52 percent were so identified for the life of the empire (1822-89). Yet on p. 86, the author states that proprietor status in this exercise is simply imputed to a large class of persons who at some period in their careers held National Guard and other government posts and honors. Lawyers in modern governments are also “linked’ with proprietors, as argued below.


Smith, Labyrinths of Power, p. 88.


Patrick L. R. Higonnet and Trevor B. Higonnet, “Class, Corruption, and Politics in the French Chamber of Deputies, 1846-1848,” in Quantitative History, D. K. Rowney and J. Q. Graham, eds. (Homewood, IL, 1969), pp. 132-145.


Derived from data in Salustiano del Campo, José Félix Tezanos, and Walter Santín, “The Spanish Political Elite: Permanency and Change,” and Hans Daalder and Joop Th. J. van den Berg, “Members of the Dutch Lower House: Pluralism and Democratization, 1848-1967,” in Does Who Governs Matter? Elite Circulation in Contemporary Societies, Moshe M. Czudnowski, ed. (DeKalb, IL, 1982), pp. 129, 227. Both studies used single coding of occupations. The Spanish data were supplied to del Campo et al. by Juan J. Linz, and do not disaggregate “other” from “unknown” occupations. The Dutch data are for occupation at the time of first election, and are given in graphical form. In the same graph, Daalder and van den Berg provide data for first occupation, showing an average of approximately 20 percent property owners among the parliamentarians from 1848 to 1967.


Putnam's U.S. elite from the 1790s to 1940 consists of presidents, vice-presidents, and cabinet members, plus U.S. senators and public officials listed in the Dictionary of American Biography. Of U.S. senators alone, however, in 1900 45 percent were in business or agriculture and 37 percent were so engaged in 1940. See Putnam, Comparative Study, pp. 185, 187, and 188.


Harold D. Lasswell, Daniel Lerner, and C. Easton Rothwell, The Comparative Study of Elites: An Introduction and Bibliography (Stanford, 1952), p. 30. (Multiple occupations were coded.)


In a study of Brazilian federal deputies who served from 1945 to 1975, David Fleischer found that 25 percent had their principal occupations in agriculture, commerce, business, industry, and transport. Although it may be assumed that a large majority of these deputies were proprietors, Fleischer notes in his text that some of those in industry and transport were workers. Therefore it is not clear how many of the 25 percent were actually owners. See Fleischer, “Thirty Years of Legislative Recruitment in Brazil,” pp. 21, 37.


Data from Cantón, El parlamento argentino, p. 40; José Luis de Imaz, Los que mandan (Those Who Rule), Carlos A. Astiz, trans. (Albany, 1970), p. 27.


The “parliamentary” subset included the following posts: president, state senate; president, state chamber of deputies (or constituent assembly); majority leader, state chamber; vice-president, federal senate; president, federal chamber of deputies (or constituent assembly); leader, state delegation, federal chamber; majority leader, federal chamber.


John D. Nagle, System and Succession: The Social Bases of Political Elite Recruitment (Austin, 1977), pp. 233, 248-249; Ralph Miliband, The State in Capitalist Society (New York, 1969), p. 66.


Miliband, State in Capitalist Society, p. 68. The analysis below, almost by definition, generally follows Miliband’s “instrumentalist” approach to political power, as opposed to the “structuralist” approach of Louis Althusser and Nicos Poulantzas. In the words of Richard W. Miller, structuralists study “the systems of institutions and ideologies through which a capitalist state fulfills its function of coordinating and maintaining capitalist relations of production in the territory in question,” while instrumentalists “tend to study the work situations, patterns of personal interaction, and sources of personal influence that produce the tendency for government officials to act in the interests of the bourgeoisie.” Analyzing Marx: Morality, Power and History (Princeton, 1984), pp. 129-130. The two approaches are complementary, as Miller implies (p. 132).


C. Wright Mills, The Power Elite (New York, 1959).


Karl Kautsky, The Social Revolution, A. M. Simons and M. W. Simons, trans. (Chicago, 1903), p. 29; Max Weber, “Politics as a Vocation,” in From Max Weber: Essays in Sociology, H. H. Gerth and C. Wright Mills, eds. (New York, 1958), p. 85-87; Miliband, State in Capitalist Society, p. 55.


In a strict sense, neither Smith nor we study the power elite as Mills conceived it (Power Elite, p. 6), for in both studies the military is not considered as a separate group from the political executive. Mills was studying the U.S. power elite at the national level in the postwar era; he considered the military establishment an independent actor, whereas it was not in the case at hand. In São Paulo, the armed forces were subordinated to civilian authority during the period in question, with the debatable exception of the Vargas years. Military personnel in the São Paulo elite amounted to 2.5 percent of the whole. See Smith, Labyrinths of Power, p. 214-216; McDonough, Power and Ideology, p. 83. Though Smith’s 1900-10 elite only included 10 percent who were property owners, and data are insufficient to conduct a test, he suspects the existence of a power elite during the Porfiriato. Yet the evidence he cites on this point only refers to one of Mexico’s 27 states in 1910 San Luis Potosí.


René Armand Dreifuss, 1964; A conquista do estado: Ação política, poder e golpe de classe (Petrópolis, 1981).


Zélia Maria Cardoso de Mello, Metamorfoses da riqueza: São Paulo, 1845-1895 (Sao Paulo, 1985), pp. 131-136. Her sampling consisted of all estates in the Primeiro Ofício da Família de São Paulo (p. 30). In three cases, the estate examined was that of the wife. “Relative” is defined here as a lineal or collateral relation through first cousin, consanguineous or affinitive. The “in-law” relationship was an important element in the dynamics of the Brazilian family.


British Chamber of Commerce of São Paulo and Southern Brazil, Personalidades no Brasil: Men of Affairs in Brazil (São Paulo, n.d.). Though this work purports to be national in scope, it focuses on São Paulo; Sao Paulo’s revolt against the Vargas government in 1932 interfered with data collection in Brazil as a whole (p. 5). It therefore would not be a reliable guide to business elites in other states. Like other biographical guides, Men of Affairs does not provide lists of assets for the biographees, and, necessarily, it is somewhat impressionistic. The fact that it was. compiled for the local British Chamber of Commerce may also bias the list toward men associated with export activities, but this was the most lucrative sector for most of the period covered. In any event, some industrialists with no listed foreign connections appear in it.


Mills, Power Elite, p. 19 and passim; Smith, Labyrinths of Power, pp. 195ff.


Some professionals may have been included in Men of Affairs simply because they were prominent in their fields. It seems more likely, however, that the British Chamber of Commerce would select professionals prominent in business circles, that is, those who had frequent contact with businessmen and who served their interests. Support for this assumption comes from the fact that, of the 36 elite members in this group, nearly half (47.2 percent) had some connection with property: either they were themselves property owners (although not so identified in Men of Affairs) or they were related to political elite members whom we know to have been property owners. This is a conservative estimate, because it is likely that other members in this group had property-owning relatives outside the political elite. It should be noted that Mills himself included professionals such as corporate lawyers in his economic elite (Power Elite, p. 232).


Mills noted that 30 of the 53 men and women in the uppermost echelon of the Eisenhower administration were “closely linked . . . with the corporate world,” ibid., p. 232. (Mills does not say, however, that 57 percent were property owners.)


Clearly, such a ranking is subject to debate. Executive committees have been ranked ahead of all other state posts except the governor, and their importance in supplying holders of other posts seems to justify this procedure. See Love, “Um segmento da elite política brasileira era perspectiva comparativa,” in A revolução de 30: Seminário internacional (Brasília, 1982), pp. 59-64. The governor was always the key figure in the political system at the state level, and in São Paulo he held a crucial post in making the national political system work. Supreme court justices are ranked at level 5 for their prestige and relative independence of the executive for the period in question.


On the gamma coefficient, see Smith, Labyrinths of Power, pp. 107-108, and for an extended discussion, John H. Mueller, Karl F. Schuessler, and Herbert L. Costner, Statistical Reasoning in Sociology (Boston, 2d ed., 1970), pp. 279-292. Statistical significance at the 0.05 level (the test used here) refers to the fact that the chances are at least 19 out of 20 that the relationship between two variables is not a random one. The appropriate significance test is for Kendall’s Tau B.


Again the significance test is for Kendall's Tau B. Residual groups—the noneconomic elite—had negative scores equal to those of the sets listed. For example, political elite members living in 1932 who were not in group 1 had a score of-.43 on NUMBER OF OFFICES and -.32 on HIGHEST OFFICE.


To confirm yet further that this relationship was not due to chance, we performed regression analyses with HIGHEST OFFICE and NUMBER OF OFFICES as dependent variables, and with an independent variable composed of the groups in Table 3. The results for both dependent variables were statistically significant at the .05 level.


See Love, São Paulo, pp. 206 and 344, n. 95, for the elite members, the foreign firms they were associated with, and the sources. All were investors or managers (or both), and not just legal counsels for these foreign companies.


McDonough points out that success in the Brazilian business world in 1972-73 still did not necessitate higher education (Power and Ideology, p. 73)—if, presumably, one was the owners son.


Miliband, State in Capitalist Society, p. 55.


Data after 1912 in Argentina or after the decade 1910-19 in Mexico are not comparable because of the compulsory exercise of the franchise in the former country and the expansion of electoral participation under the revolutionary governments of the latter. The relevant issue, therefore, is not statutory universal male suffrage, which had existed in several Platine provinces in the early nineteenth century, but mass political participation, which only entered Argentina after the obligatory suffrage of the Sáenz Peña Law of 1912.

In an elite study examining shares of the Argentine Chamber of Deputies held by “aristocrats” and “nonaristocrats,” Peter H. Smith has shown that the proportion of aristocrats in that body fell from over 60 percent in 1904-15 to 30-35 percent in the mid-1920s, as one might expect as a consequence of the effective expansion of suffrage. For our purposes, however, it should be noted that not all of Smith’s aristocrats appear to be property owners. In any case, his share of aristocrats does not correspond very well to that of Cantón’s property owners in the Argentine Chamber of Deputies. See Smith, Argentina and the Failure of Democracy: Conflict among Political Elites 1904-1955 (Madison, WI, 1974), pp. 24, 26-27, 118-122; Cantón, El parlamento argentino, p. 40.


Smith, Labyrinths of Power, p. 216.


These years included periods of rapid inflation, so it makes a difference what deflator is used to measure income and outlay in constant terms. Using three price indices, we determined that São Paulo’s revenues in real terms rose between 3 and 3.5 times from 1891 to 1892. Using the same deflators, we found that central government revenues per capita declined between 1882-89 and 1892-99. With two price indices, Nathaniel Leff found that central government expenditures per capita also declined from the decade 1882-91 to 1892-1901. See Love, São Paulo, p. 303; Leff, Underdevelopment and Development in Brazil, 2 vols. (Boston, 1982), I: Economic Structure and Change, 1822-1947, 244-246 (price indices); II: Reassessing the Obstacles to Economic Development, 106; Instituto Brasileiro de Geografía e Estatística, Anuário estatístico do Brasil: Ano V 1939/1940 (Rio de Janeiro, n.d.), p. 1,410 (federal revenues and expenditures); and O Brasil em números (Bio de Janeiro, 1960), p. 5 (population).


Love, São Paulo, p. 248.


Ibid., p. 71.


Nicos Poulantzas, Political Power and Social Classes, Timothy O'Hagan, trans. (London, 1973), pp. 346-347. The ruling class, however, had only conditional ownership until the early modern period.


See Perry Anderson, Lineages of the Absolutist State (London, 1974), pp. 25-26; Tilomas H. Holloway, Immigrants on the Land: Coffee and Society in São Paulo, 1886-1934 (Chapel Hill, 1980), pp. 114-121; and Love, São Paulo, p. 129.


At least for ministers (Carvalho, Construção da ordem, p. 79).


For Marcello Carmagnani, the feudal structure of Latin America was dominated by a merchant monopoly system in which unequal exchange (in terms of labor costs) for goods produced 011 opposite shores of the Atlantic was combined with extra-economic compulsion in Ibero-American labor organization. In this scheme, only the rise of U.S. monopoly capital circa 1914 began the decline (still in process) of the feudal economy, since U.S. capital— unlike its British predecessor—began to introduce capitalism within the domestic economies of Latin America through direct investment in industry. Carmagnani, Formación y crisis de un sistema feudal: América Latina del siglo XVI a nuestros días (Mexico City, 3d ed., 1980), pp. 37-40, 157, 164.


Holloway, Immigrants on the Land, p. 171; Love, São Paulo, p. 4.


Love, São Paulo, pp. 79, 153, 156-158, 169, 171, 174.


Verena Stolcke and Michael M. Hall, “The Introduction of Free Labour on São Paulo Coffee Plantations,” The Journal of Peasant Studies, 10:2-3 (Jan.-Apr. 1983), 183.


Holloway, Immigrants on the Land, pp. 169-170. Admittedly, not all students agree with this interpretation. One who does not is José de Souza Martins, who argues that the colono system was not fully capitalist, because the usufruct feature did not allow a complete separation of the worker from the means of production—a condition Marx specified as unique to capitalism and its wage labor system. Yet Souza Martins still views the fazendeiro class as capitalist.

We prefer the view of Stolcke and Hall, who see the usufruct system as a means of minimizing planters’ cash outlays and maximizing their profits. This situation occurred because prices of food were initially high, much of it being imported from other states; because usufruct would attract immigrants and therefore contribute to labor market saturation, forcing down wages; and because peak labor demands for coffee and food crops did not occur in the same season. See Souza Martins, O cativeiro da terra (São Paulo, 2d ed., 1981), pp. 16, 91 (on fazendeiros as capitalists); pp. 18-19 (on noncapitalist features of the colonato); and Stolcke and Hall, “The Introduction of Free Labour,” 184.


Stolcke and Hall, “Introduction of Free Labour,” 185-186.


Love, São Paulo, p. 76, and Holloway, Immigrants on the Land, p. 98.


Holloway, Immigrants on the Land, p. 130.


It is true that planters systematically employed extra-economic means of compulsion to despoil native-born peasants, not to mention indigenous peoples, of their lands on the coffee frontier. Although this “primitive accumulation” accompanied the westward march of coffee, there were no significant precapitalist forms of labor control within the coffee economy. On violence on the frontier, see Love, São Paulo, pp. 15, 70-74.


E. P. Thompson, “The Peculiarities of the English,” in The Poverty of Theory and Other Essays (New York, 1978), pp. 251-252, and 298, note 13, citing Marx: “. . . in the ‘categoric’ sense the farmer is an industrial capitalist as much as the manufacturer.” A non-Marxist, Fernand Braudel, insists that plantations were “capitalist creations par excellence.” Civilization and Capitalism: 15th-18th Century [sic], II: The Wheels of Commerce, Siân Reynolds, trans. (New York, 1982), 272.


There is a large Marxist literature, beginning with Marx and Engels, on the conditions under which the state can be “relatively autonomous” of the ruling class. See a discussion of this literature in Poulantzas, Political Power, pp. 253—321. In Brazil, however, the most celebrated thesis about the relative autonomy of the state is set in a Weberian framework; see Raymundo Faoro, Os donos do poder: Formação do patronato politico brasileiro, 2 vols. (Rio de Janeiro, 2d revised ed., 1975). Two Marxists who argue for the relative autonomy of the state are Otávio Guilherme Velho and Joe Foweraker. See Velho, Capitalismo autoritário e campesinato (Um estudo comparativo a partir da fronteira em movimento) (São Paulo, 2d ed., 1979); Foweraker, The Struggle for Land: A Political Economy of the Pioneer Frontier in Brazil from 1930 to the Present Day (Cambridge, 1981).


This view is not meant to imply that Sao Paulo had a “conquering bourgeosie,” in the phrase of Charles Morazé; on the contrary, it is clear that the Paulista bourgeoisie was dependent on international capitalist structures, signals, and inputs. See Love, São Paulo, pp. 203-211.


Topik, “The State’s Contribution,” pp. 217, 226-228; and Font, “Padrões de ação coletiva,” pp. 218-244.


Font, “Padrões de ação coletiva,” pp. 219, 242, 244. Elsewhere, Font cites Gramsci as the source for his views on hegemony. See Font, “Planters and the State: The Pursuit of Hegemony in São Paulo, Brazil (1889-1930)” (Ph.D. diss., University of Michigan, 1983), pp. 13-14.

It is worth recalling, as Winston Fritsch has in this context, that Gramsci himself allowed for concessions hy the hegemonic group. Gramsci asserted that “. . . a certain balance of compromises [is presupposed in the notion of hegemony, in which] the leading group makes some sacrifices of an economic-corporative kind.” Fritsch believes that coffee interests were hegemonic in Gramsci’s sense. See Fritsch, “Aspects of Brazilian Economic-Policy under the First Republic (1889-1930)” (Ph.D. diss., Cambridge University, 1983), pp. 247, 248, 253; Antonio Gramsci, The Modern Prince and Other Writings (New York, 1970), pp. 154-155.


Francisco C. Weffort, “Classes populares e política (Contribuiçao ao estudo do ‘populismo’)” (Ph.D. diss., Universidade de São Paulo, 1968), p. 73. The idea is developed in Boris Fausto, A revolução de 1930: Historiografia e história (São Paulo, 1970).


Font, “Padrões de ação coletiva,” pp. 221, 224, 243.


On these points, Font elaborates on the previous literature. See Love, São Paulo, pp. 225-227 (disputes among planters), and Holloway, Immigrants on the Land, chap. 6 (immigrants as landowners).


By decade, the percentages of fazendeiros among total elite members were 46 percent (1889-99); 51 percent (1900-1909); 51 percent (1910-19); and 35 percent (1920-29). The Partido Democrático executive committee was excluded in this exercise because Font’s thesis refers to PRP governments, from which the PD was completely excluded. The latter, in any case, had a smaller share of fazendeiros on its executive committee than did the PRP (35 percent vs. 60 percent) over the lifetime of the PD (1926-34). Love, São Paulo, p. 165.


These intraparty battles are spelled out in Love, São Paulo, pp. 221-225.


For example, Júlio Mesquita, publisher of O Estado de S. Paulo and a dissident in 1901 and 1915; Paulo de Morais Barros, in 1901 and 1915; Alfredo Pujol, in 1901 and 1915; and Paulo Nogueira Filho, in 1915.


Love, São Paulo, pp. 117, 163-166.


Ibid., pp. 248, 250, 253. By contrast, manufacturers, whose power, according to Font, was rising in the ’20s relative to that of planters, benefited negligibly from government largesse. The only instance of a state subsidy to a manufacturer in the decade amounted to 0.4 percent of the budget in 1929, and was not renewed. Ibid., pp. 257, 302, 304.


In fact, magnate planters by the 1920s had diversified their holdings to include investments in banking, commerce, industry, and exporting. Thus, the existence of overlapping portfolios tends to muddy, if not invalidate, any thesis of sectoral clash within the Paulista economy. Love, São Paulo, p. 67.


Fritsch, “Aspects of Brazilian Economic Policy.” p. 197.


Gunnar Myrdal, The Challenge of World Poverty: A World Anti-Poverty Program in Outline (New York, 1970), chap. 7: “The Soft State,” pp. 208-252.


Love, São Paulo, p. 263. The Reajustamento Econômico was partly conceived as a device to compensate planters for an exchange rate that would no longer favor their interests. Thus, it appears that, in the view of the Vargas government, as against Topik’s opinion, planters had enjoyed an advantageous exchange policy up to that point. See Alberto Venâncio Filho, A intervenção do estado no domínio econômico: O direito público econômico no Brasil (Rio de Janeiro, 1968), pp. 114-115, and Annibai Villanova Villela and Wilson Suzigan, Politica do governo e crescimento da economia brasileira, 1889-1945 (Rio de Janeiro, 1973), pp. 200-201.


The relative autonomy of the state is inferred from Topik’s finding that both the “export oligarchy” and other economic groups did not dominate the state, and from the judgment, “People in power believed that they had to represent the long-run interests of the nation, not just the short-run interests of the export oligarchy.” Topik, “The State’s Contribution,” 227-228.