At the end of 1989, Brazilians voted in direct presidential elections for the first time since the military seized power in 1964. After an inconclusive first round, victory in the second went to Fernando Collor, an independent conservative, by a small margin over Luis Inacio da Silva, universally known as ‘Lula’, the leader of the Workers’ Party (Partido dos Trabalhadores) founded only ten years earlier. Collor took office in March 1990, and completed his first eighteen months in power in the midst of a deep economic and political crisis which continues today. In this article I seek to place the present moment in historical context. Over a quarter of a century ago, on the eve of military intervention, Brazilian sociologist Octavio Ianni saw Brazil as poised between a developing but dependent capitalism (as a potential junior partner of international capitalism), and socialism. He hoped that despite the immaturity of the proletariat, it might take advantage of divisions between the industrial and agrarian bourgeoisies to take the political future of the country into its own hands.footnote1 He traced the character of Brazilian politics to the failure of the industrial bourgeoisie which appeared under state tutelage after 1930 to consolidate itself as a united class. As a result,

power continually oscillates this way and that, requiring clever, conciliatory, flexible, malicious leaders. The speciality of some Presidents, especially Getulio Vargas (1930–1945 and 1950–1954) and João Goulart (1961–1964), has been a talent for half-tones, for nuances, for duplicity, for continual manoeuvring of parties, factions and classes. It has been a singular Prince who has ruled in Brazil.footnote2

Before Ianni’s analysis even reached the pages of nlr, the 1964 coup removed any early prospect of socialism, and set Brazil firmly on the path of dependent capitalism. With the military dictatorship now over, and a new and singular prince in power, it seems that again a socialist option, this time far more palpable than any available in 1964, has been rejected in favour of continued adherence to the harsh disciplines of dependent capitalism in its contemporary guise.

Since 1964 Brazil’s industrial growth, along with rapidly expanding agribusiness and large-scale investment in mining in the Amazonian basin, has made it the eighth largest economy in the world. Despite its extreme concentration of wealth and income, its population of close on 150 million gives it a substantial internal market. Economic policy under the military sought to attract foreign and domestic capital into large-scale capital-intensive activity supplying consumer goods such as cars, televisions, hi-fi equipment and computers to the top end of the domestic market, and goods ranging from processed or unprocessed primary commodities (soya beans and meal, orange juice, coffee, iron ore and pellets) to manufactures (such as steel, cars and commercial passenger aircraft) to the world market. The state played a key part in underpinning the ‘economic miracle’ which peaked between 1967 and 1974 through massive investment in infrastructure and basic industrial production (notably in steel, hydro-electricity and telecommunications), direct and indirect subsidies to private capital, and the exclusion and repressive control of the majority of the population. This ‘savage capitalism’, seeking headlong growth regardless of the social cost, has transformed the social structure of the country, producing an urban and rural proletariat on a scale unimaginable a quarter of a century ago.footnote3 The model of economic development has rested upon a ‘triple alliance’ between the Brazilian state and foreign and domestic capital, and relied heavily upon foreign borrowing, making Brazil the largest Third World debtor in the 1980s. Even on its own terms, the extent of success has been limited. As Arrighi notes (in nlr 189), the economic crisis that prevailed throughout the 1980s returned Brazil to the ‘historic level’ of 12 per cent of the gnp per capita of the organic core of the capitalist world economy at which it had been stuck between 1938 and 1970. In the process, inequalities of wealth and income within and between regions sharply worsened, and in urban and rural areas alike open conflicts between classes are more evident today than in the period of alleged anarchy which preceded military takeover, while radical projects are more widely canvassed and new patterns of independent organization and mobilization have multiplied.

In one respect, though, nothing has changed. The return to democracy has been marked by a deepening institutional crisis which again underlines the perennial incapacity of the dominant classes in Brazil to protect and further their interests within a democratic regime. This incapacity cannot be understood without delving back deeply into Brazilian history, as it dates back not twenty-five years but all the way to the founding of the Republic a century ago. Equally, it requires close attention to the complex relationship between Brazil’s underlying political economy and the pattern of discontinuous institutional change to which it has given rise. In this article I trace it back to the political arrangements of the Old Republic (1889–1930), and the relationship there between clientelism and class power. I then follow it through the periods of populism (1930–1964) and military rule (1964–1985) into the first transitional phase of the New Republic (1985–1989). This leads into an interpretation of the Collor presidency, and the challenge represented by the Workers’ Party. I argue that successive episodes of regime breakdown reflect contradictions between the state, forms of institutionalized class representation, and Brazil’s changing internal political economy. In the recent transition the chronic inability of the dominant classes to organize themselves politically has been compounded by the legacy of a politics of state clientelism practised by the military. Collor appeared as a deus ex machina when this proved bankrupt, but his appearance was more a reflection of the deep crisis afflicting the dominant classes than a solution to it. As a consequence, he has perpetuated the weakness of those classes where democratic politics is concerned. He came to power without an organized party behind him, and confronting as a major adversary an independent workers’ party based in the most dynamic centre of modern industrial activity in. the country. At the same time, the Workers’ Party has not established itself unassailably as the leading opposition force in the country. With only 35 seats won in the Chamber of Deputies in the elections of October 1990, it appears as one of a number of left or centre-left alternatives, competing with the Democratic Labour Party (pdt) of Leonel Brizola (46 seats), and the Social Democratic Party (psdb, 42 seats), in which Mario Covas, presidential candidate in 1989, and Fernando Henrique Cardoso, a leading intellectual formerly in the pmdb, are major figures. One purpose of this article is to place these forces in historical context, as a prelude to a later assessment of possible strategies for the Left in general.

After a brief period of military control (1889–1894), the political system of the Old Republic, which endured until 1930, passed into civilian hands. It was largely dominated by the leading states in the federal system (São Paulo, Minas Gerais, and Rio Grande do Sul); export-oriented elites broadly committed to a laissez faire system internationally and at home were the major beneficiaries. The relatively orderly passage to the Republic, bringing an end to an empire that had lasted for sixty-seven years, was a remarkable achievement. The empire had been a centralized regime in which loosely coordinated national parties alternated in power at the behest of the emperor himself. It had relied upon slave-labour for the production of coffee for export, and for much activity throughout the domestic economy, while its social base had centred on established landed elites in the old sugar heartlands in the north-east, and the first of the coffee provinces, Rio de Janeiro. The system was clearly in crisis from the late 1860s, but did not collapse until the final abolition of slavery in 1888 was swiftly followed by a bloodless military coup in November 1889. Within a decade, an entirely new political system had been established. Centralization gave way to decentralization, with federalism instituted by the Constitution of 1891; slavery gave way to free labour; and political dominance passed from the slavocrats of Pernambuco, Bahia and Rio de Janeiro provinces to the republican landowners of the state of São Paulo, in partnership by the turn of the century with the leaders of neighbouring Minas Gerais. The loosely organized national parties of the Empire gave way to parties constituted at state level, and elections, in which only a tiny minority of the adult male population took part, fell under the control of local elites and statelevel political machines.

The politics of the period have generally been seen as dominated by the practice of coronelismo, in which landowners (known as ‘colonels’ (coroneis) because of their commonly held honorary rank in the National Guard) acted as patrons at local level, and bartered their ability to deliver the vote for recognition as local powerholders by the state. In Nunes Leal’s classic description