5 The Economics of Slavery

Slaves are the hands and feet of the senhor do engenho for without them Brazil could not make, conserve, or expand its resources, nor have a viable plantation system.¹

This famous affirmation of Padre Antonil made at the beginning of the eighteenth century could not be more true and need not be restricted to the senhores de engenho (or sugar mill planters). They were the basic labor for the majority, even of the free population, which depended on their services. In all the economic activities developed in Brazil until the last quarter of the nineteenth century, slaves were the essential labor force.

At the same time that the importance of slaves was recognized, there was also a biased vision created of that contemporaneous society in Brazil. This famous phrase, despite its obvious truth, perhaps lead to a distortion of the Brazilian society of the period. The idea of a slave mass and a planter elite to the exclusion of all other classes dominated Brazilian historiography until the first half of the twentieth century. Plantations, monoculture, and slave labor were considered the essence of Brazilian society. This society was segmented between just slave owners and slaves, with the latter seen as “things,” culturally backward, morally perverted, without any contribution to Brazilian culture. Fortunately, this vision of Brazilian society began to change beginning in the 1930s, particularly with the pioneer work of Gilberto Freyre, who gave life to the slaves and showed their important contribution to the formation of this culturally mixed society.

But if Freyre and his predecessors in the Bahian Afro-Brazilian movement finally changed the relative weight of Africans and their descendants in the formation of modern Brazil, they still maintained the older idea of a society filled with just planters and slaves. However, when we talk of masters, we should qualify what masters we are talking about. Usually the senhor was equated with slave owner. But who were Brazil's slave owners? What were the characteristics of the slave owners and in what activities did they use their slaves? We can give dozens of types of slave owners with social positions and economic interests completely different from one another. When we speak of a senhor de engenho, we are speaking of an economic and social elite, with dozens or even hundreds of slaves, located in the sugar cane agriculture in the northeast, in Rio de Janeiro, or in São Paulo. This type of elite was also crucial in the large coffee plantations (fazendas) in the Paraíba Valley in the last century of slavery. Perhaps large-scale mining in Minas Gerais in the so-called lavras presents similar characteristics. These are activities on a large scale with groups of slave workers organized in gangs with little specialization but with great physical labor. The rigid control and the physical violence needed to be productive perhaps are the factors that mark this form of economic exploitation. Nevertheless, one asks, how representative were these activities in Brazilian society? They were, obviously, economically very important, because they represented the principal economic activities in various periods of Brazilian history and were the reason for its occupation and colonization and the form of its insertion into the international market, within the rules of the colonial system or as an independent nation. But we cannot characterize slavery in Brazil only by the standards of these activities. At least from 1700, at no time in the history of slavery in Brazil were the slaves of the engenhos, the mining camps, or the extensive coffee fields the majority of the slaves then residing in Brazil. They were mostly in fact owned by slave owners quite different from the big planters and miners. Can we classify in the same economic and social category the priest who possessed on aged female slave who aided him in his parish work? Or the freed black woman who declared that she was poor in eighteenth-century Minas Gerais, but possessed seven slaves? Or the thousands of farmers who developed an agriculture directed toward the internal market and also produced subsistence crops who had one or two slaves to assist the farmer's family in their small plots? How do we define the senhor who used his slaves in commerce in the towns, or the itinerant miner (faiscador) who owned some slaves who were given complete freedom of movement and paid a daily wage in gold? How do we characterize the owners of coartado slaves (self-purchasing slaves) who had achieved an effective autonomy and who gradually paid for their own freedom? How do we characterize a farmer who possessed one slave and practiced family farming, working on a small plot with his slave? How do we characterize the master who liberated his slave, but conditions his manumission on the continued service to the master until his death? How does one classify the relations of the slave owner with his illegitimate child still maintained in slavery? Finally, how to place the agregado (servant) who resides in the home of others, has no identifiable activity, and owns slaves? These suggest a far more complex stratification and social organization than was originally suggested by the discussion of slavery in Brazil. The abundant current historiography on slavery shows everyday situations more complex, although equally frequent, which defined slavery in Brazil.

On the other hand, the category of slave is well determined, both legally and in practice, in Brazil and needs not be questioned. However, when we study the effective social space of slaves in Brazil, we find a variety of situations that need to be clarified in order to understand the organization and functioning of slavery in Brazil. Let us give some examples. How do we compare the social role, opportunities, and freedom of action for slaves in production compared to the enormous number of slaves in domestic service? How does one compare these three factors for slaves in the urban or rural area; for slaves in agricultural production and skilled slaves who transformed cane into sugar or were skilled artisans? How does one classify a slave married to a free person, or a slave whose children are free? Or, on the contrary, what is the relationship of a slave to his owner when the owner is a forro (or liberated slave) or is a master who lives from alms? How does one classify a household composed just of slaves; where does one put the coartado slave? What is the social position of a slave who was a child of his owner, and was recognized and freed in the owner's will? How does one compare the slaves who sweated in the gangs of the large agricultural estates with the slaves involved in the management of the mule trains or who raised cattle? These are just a few examples that suggest the complexity of slave society in Brazil and the varieties of social space that the slaves could possess within the system. The recent studies on Brazilian slavery show such extensive diversity among slaves and their masters and permit one to better understand the complex social relations prevalent in Brazil during the time when slavery existed in the colonial and independence periods.

The studies on slavery in Brazil, which have been carried out in the last thirty years, have revealed a very complex society from the social and economic perspective. We find manumitted slaves everywhere, and even some who owned slaves. In the diamond district of Serro do Frio at the height of the mining boom, they represented an extraordinary 20 percent of the slave owners. Would it be possible to classify them as senhores? Would they have the social status of a senhor? We have also found slaves in all kinds of activity. They are the labor force in all the large mines and plantations, but they are also in foodstuff production, artisanal activities, in commerce, in transport, in fishing, in art, and of course, in domestic labor. They are more restricted in the rural areas and freer in the cities. Meanwhile, miscegenation continues nonstop and soon the mestiços, mulatos, and pardos are the majority of society – as slaves, liberated slaves, and free persons of color. In spite of these qualifications, it is important to stress that the slave regime in Brazil, despite its complexities and differences in time and space, was based on violence, apparent or not, which was the only way such a system of labor could survive in any society. To show its more complex nature than previously understood, does not soften the violence of the slave regime but is rather to better understand that society. As one famous 1839 Brazilian Manual for Agriculturalists warned its readers, given that slavery was a violation of natural rights, the only way to force slaves to work was to instill “fear, and only fear, but employed with system and art, because excessive fear will be counterproductive.…”²

The analysis of an interpretive model of slavery in Brazil is not an easy task. Our interest is to find better means to classify these economic and social relations, without preoccupying ourselves with a theoretical analysis of the modes of production, which was a major theme in the middle of the twentieth century.³ We thus have examined the available literature for models to help us understand in general what was occurring without losing the specificity of the Brazilian experience or oversimplifying the Brazilian reality. The economist Stefano Fenoaltea has developed an explicit model of a slave economy based on both classical and modern slavery.⁴ He affirms that in tasks that required greater physical effort, little skill, and little need for careful work, physical violence and the need for close supervision are the defining characteristics. In these cases, better output could be obtained from slaves by the effective or potential threat of violence. The costs of use of violence and strong supervision are compensated for by greater output. This was characteristic of the forms of labor usage in the great mines and plantations. On the other hand, when the work demands careful attention, dedication, and skill, the use of force is not as efficacious. In these cases, various forms of positive incentives have to be adopted. The Fenoltea model can be used to define the different labor systems used in Brazilian slavery. Routinized gang labor with actual and potential use of violence as the dominant form of labor control clearly prevailed on the large plantations and mines. In all of the other extraordinarily varied activities in Brazil, there was in fact a mix of both positive and negative incentives. It was these positive incentives that gave raise to the manumitted slaves, both voluntary and self-purchased. Everyone from slaves in households through itinerant venders and miners and day laborers had access to income or other positive forms of incentive to encourage their labor. In some activities, it was the most efficacious means of obtaining better results from slave labor. Humanitarian reasons, even in the case of their illegitimate children, although present, were not the dominant reasons for the granting of voluntary manumissions. Moreover, even in tasks where positive incentives were used, violence was present. The master had absolute power over the slave and could inflict serious physical punishment. As Taunay argued, the masters who were planters should exercise their powers like “heads of small kingdoms, in which they should govern despotically, and take upon themselves the attributes of legislator, magistrate, captain, judge, and sometimes even executioner.”⁵ The master controlled the freedom of the slave and of the slave's child if he or she was born of a slave mother. Often a slave obtained his or her freedom, but not the freedom of his or her spouse, children, or parents. Violence was also implicit in the system that allowed families to be separated by the sale of its members. Thus, even with positive incentives, it is not possible to talk of slavery without recognizing its inherit violence.

We could give an example that clearly illustrated the two types of labor, even in the same occupation. The extraction of gold and diamonds occurred in Minas Gerais in the first half of the eighteenth century. However tight and vigilant was the supervision of the slaves, it was still necessary to count on their responsibility and initiative in the areas of both gold and diamonds. For this reason, the miners tried to stimulate their slaves in various ways, even in the large mine works. In this respect, we can separate two phases. In this brutal phase, which demanded great physical effort, slaves were used to move soil and rocks on a large scale to open up the veins of gold and diamonds. This task demanded control and supervision but not intense attention on the part of slave drivers. The other aspect, the panning for gold and diamonds, also demanded control, but this more violent method was not sufficient. It needed the interest, care, and attention of the slave workers themselves. The extraordinary watercolor paintings and lithographs of Rugendas and Carlos Julião show the rigor of control in this phase, with each slave having his or her own controller. This unusual one-on-one supervision guaranteed that slaves could not steal the gold or diamonds, but did not prevent them from throwing away valuable minerals. In this final phase of mining, it was normal for slave owners to offer all types of rewards, especially in the extraction of diamonds. Finding a precious stone could be worth even manumission. In the case of miners with few slaves and little capital, a whole range of incentives and arrangements were worked out between the master and his or her slaves. It was usual, for example, for the slave faiscador (itinerant miner) to be given total autonomy in his prospecting and to pay a daily fixed wage in gold oitavas to his or her owner.⁶

By saying that most forms of slave labor were based on positive incentives and not on direct physical violence, we are not saying that violence was not one of the basic characteristics of modern slave labor that developed in Brazil. Clearly violence was behind the original capture of free persons in Africa and bringing them against their will to America. The violence of the capture, the Middle Passage, and the forced abandonment of families and friends were endogenous to the whole system. Recent studies have shown the many indications of this violence and the individual and collective manifestations in terms of revolt and aggression. But the society that was formed in Brazil was sufficiently complex to generate another type of violence similar to persons born free: violence of a personal nature, the fruit of social relations that were established there. This is seen in violence between slaves themselves and against free persons who were not their masters. Like violence in any human society, these reflected social conflicts or human passions. Sometimes it is difficult to separate these two types of violence in a slave society, but they should be separated. Clearly, the violence directed against the slave to extract labor and the violence expressed by the slave against the system and his or her master is one pattern of violence, and the daily conflicts among equals is another. But it is also crucial to realize that there were positive incentives that made total economic sense, especially in those tasks demanded by the slave owners in which violence would have been counterproductive and economically costly.

This discussion of positive and negative incentives was a fundamental part of the argument of economists Robert Fogel and Stanley Engerman in their analysis of North American slavery.⁷ However, their stress was on a supposed uniform level of positive incentives with a minimum of violence across all slave activities. For this, they were criticized by a large number of scholars who showed just how important close supervision and violence were in the productivity of southern U.S. slave labor.⁸ What is important about the Fenoltea model is that it recognizes, on the basis of studying numerous slave systems, that different types of slave labor can be quite different in terms of productivity on the basis of either positive and negative incentives. Clearly recognizing these differences goes a long way toward explaining the surprising variations to be found in Brazilian slavery, and even the very early and very rapid expansion of the free colored class long before the final abolition of slavery in 1888.

Another factor worth considering in analyzing the economic structure of Brazilian slavery is the surprising distribution of slaves among all classes and groups of individuals, from elite white planters to poor free colored women. This is a pattern distinctly different from that which developed in the United States despite the fact that these slave systems look alike in relative importance of the slaves and slave owners within each society and the relatively modest size of slaveholdings compared to the West Indies. But as the economist Gavin Wright has suggested, there was a sharp difference in the United States among agricultural producers in terms of slaveholdings.⁹ Slavery was almost exclusively associated with a high level of commercial production, whereas the less commercial farming sector usually did not have slaves. Among small producers, the tendency was toward subsistence production of foodstuffs, with a modest placing of surplus production on the local market. The long-term high risk associated with full-time market production and the high costs of entry into this sector kept these farmers out of the export sector. For Wright, this division based on long-term risk marked the distinction between slaveholders and nonslaveholders. The former were primarily associated with the export market, and the latter were primary into local foodstuff production. He thus stresses the association between high market risk and adoption of slave labor.

Although there is no question that the export sector within Brazil absorbed a large and growing part of the slave labor force over time, there was a surprising distribution of slave ownership among food producers and even some primary subsistence farmers and a high level of slave ownership among artisans. Clearly, there was a long-term risk associated with export agriculture and mining, even when attempts were made to reduce that risk for producers. Thus, the division between mill owners and cane farmers in the early sugar industry of the northeastern region was an attempt to reduce risk. The mill owners took the greatest risks and had the highest capital investment, whereas the cane producers did not have the capital costs associated with the costly milling process. But even for the independent cane producers, there was a high level of risk based on prices in the international market, especially given the long cycle of production of sugar cane. In the second half of the nineteenth century, the great fazendeiros of coffee presented a similar risk. From the beginnings of the nineteenth century, when native gold mining was replaced by foreign capital, it too used slave labor and also involved long-term market risk due to both market and geological factors that were present for the entire period of their operations. Investments in prospecting were immense and not always compensated.

But this model of risk and slavery does not hold quite as well for Brazil in relationship to low-risk production. For four centuries, slaves were involved in both export and local market production and in all aspects of agricultural production including foodstuffs and subsistence, along with a host of artisanal services. Although slaves were part of the rural labor force, including the gangs utilized in the plantation, a large part of the slave population was allocated to domestic service subsistence agriculture, manufacturing, and artisanal services for the local market. There were also small family farmers who produced some surplus for the external market who sometimes owned slaves. Initially, this was the case in coffee and early alluvial gold mining in Minas Gerais. Moreover, despite the major development of the slave plantation model in Brazil, as in sugar for the northeast and then for Rio de Janeiro and São Paulo and in coffee in the nineteenth century in Rio de Janeiro, São Paulo, and Minas Gerais, its importance in terms of slaves assigned to its control was limited, accounting for no more than 20 percent of the resident slave population.¹⁰ Even at the height of the mining boom, the slaves involved in the large-scale lavras were only a minor part of the slave population in the mining regions. The Brazilian case shows a diversity of use of slaves well beyond the plantation. Moreover, we find nonslave-owning producers in all types of activities except in sugar milling. This diversity of slaves and nonslave producers is the principal characteristic of slavery in Brazil.

Given the lack of a positive defense of slavery in colonial and nineteenth-century Brazil,¹¹ there never emerged a school of thought that defended the institution on noneconomic grounds. The assumption in most of the historical literature has been on the economic rationality of slaveholdings. The Marxist historiographical tradition (as exemplified by the work of Cardoso, Ianni, Caio Prado, and most recently Jacob Gorender) has taken the classical liberal position on the inefficiency of slave labor and its supposed incompatibility with modern capitalism but has not questioned its economic basis.¹² Because of this tradition, there have been relatively few detailed studies of the economics of the institution. The best such studies, which grew out of the classic work of Conrad and Meyer,¹³ have been those of Pedro Carvalho de Mello,¹⁴ Flávio Versiani,¹⁵ Antonio Barros de Castro,¹⁶ Iraci del Nero da Costa,¹⁷ and Roberto Monasterio.¹⁸ All have found that the patterns of capital investment in slave labor for coffee or sugar or even the jerked beef industry in Rio Grande do Sul were economically rational and that the relative efficiency of slave labor was competitive with free labor in the types of agricultural activity in which slaves were employed. Moreover, recent research has shown an increasing interest in economic rationality in the use of slave labor, both in Brazil and in the other nineteenth-century American slave societies, among the administrators of slave labor. Using models from free labor control and administration, slave supervisors and overseers tried to apply these techniques to their slave workers. In Brazil, these ideas first evolved in the first half of the nineteenth century,¹⁹ but achieved greater importance in the second half of the century with the great expansion of coffee plantations and especially given the scarce and increasing cost of labor. These ideas represented a modern view of labor administration and reflected a new rationality that was appearing in the capitalist world.²⁰

Although the debates on the modes of production model still find an echo in some historical schools in Brazil, there is another aspect to this debate aside from the classic Marxian formulation. This is the issue of whether the usage of slave labor inhibited or limited the technological development in the Brazilian economy. This question is difficult to answer given the lack of studies on the subject in terms of the Brazilian experience. It is undeniable that sugar production in the sixteenth and seventeenth centuries represented the latest in advanced technology of the period. The relative decline of the Brazilian sugar industry was due to the rise of newer centers with better technology, but these new centers also used slave labor. This was especially the case in Jamaica and Cuba, which saw a revolution in sugar technology in the nineteenth century even while employing slave labor. It is true that more specialized jobs in sugar technology were usually reserved for free labor, but there was nothing to impede the planter class from using such labor on their estates, and many did so. Moreover, some of these same tasks and skills were to be found among slave workers. The economic backwardness of the Brazilian mining industry, much discussed in the literature, was not due to slave labor, but to the lack of experience among the free Portuguese and Brazilian miners. Moreover, when the latest technology was introduced by English miners in the nineteenth century in Minas, these foreign companies employed slave labor just like all other enterprises in Brazil.²¹ In the textile mills, mines, and plantations of Brazil, there were no constraints on the use of free and slave labor at the same time, although the more rudimentary and unskilled parts of this work were usually confined to slaves. Free and slave workers were used together and sometime slave workers were given as highly skilled a task as were free workers.²² The use of positive incentives in some of the more skilled tasks in the service area or in production often left the slave workers (in terms of remuneration) in a similar position as free salaried workers. Many slaves were able to obtain their freedom in return for labor or by purchasing their freedom using the salaries they obtained in these types of jobs. Assembly-style factory labor with its careful division of tasks was probably difficult to maintain with slave labor, although slaves could be found in all the early textile and other factories that were established in the nineteenth century. It is not difficult to imagine that slaves could have been used in this industrial labor force given the early period of industrialization in the nineteenth century when the labor force consisted of child and women laborers, poor working conditions, long hours, and the use of physical violence. It would thus appear that the use of slave labor was not an impediment to technological development, which was probably more influenced by capital and market conditions than by the availability of free-wage workers.

There has recently been considerable work done on the market for slaves and most particularly on the changing price structure of this market. All such price studies have also shown that the price structure of slaves represents a positive view of investors in the future of the institution, with the price of adult woman in the ages of fertility including a premium for the potential to create slave children – the so-called “positive price for unborn children.”²³ They have also shown the slow and steady evolution of a common national market for slaves in a convergence of slave prices over time. The largest such studies have been the work of Bergad on Minas Gerais,²⁴ Carvalho de Mello on Rio de Janeiro,²⁵ Katia Mattoso on Bahia,²⁶ and Versiani and Vergolino on Pernambuco.²⁷ There have also been shorter term or more local price series generated for occupations, ages, and sex by a host of other scholars.²⁸ All such studies have found the slave market to be rational and have shown that by the nineteenth century, there was a coherent national market for slaves and that slave prices moved in similar fashion in every region in Brazil, from the coastal and interior areas of Pernambuco to the coffee regions of São Paulo and the interior towns of Minas Gerais.

The maintenance of a modern slave system for four centuries required a complex organization in terms of a market for slaves, which in essence was Brazil's basic labor market. Slave prices were not fixed, but reflected variations in market conditions and varied over time and by region. Because buying a slave represented a future purchase of labor, the slave owner has to take into consideration the market conditions of his products. But what were the factors that determined the supply and demand of slaves and, consequently, of their labor? First, we should point out the conditions of the offering of slaves in the principal slave markets and the organization of the slave trade. As noted in the previous chapters, the conditions of supply were influenced by the traffic itself. Internal conflicts in the African areas furnishing slaves changed over time, along with the permanent disputes between the principal nations involved in the control of slave supplies and in their movement to America. Despite the progressive expansion of an internal market of slaves born in Brazil, the Atlantic slave trade was the determining element in the fixing of prices until 1850, particularly in the regions where the plantation export crops predominated. As we have noted, slaves were used in the most varied activities, but probably the most important economic activities and their profitability determined the demand conditions, particularly for slaves most apt for the labor needed by such activities: adults in good physical condition who also had some type of desirable specialization. Besides, as the slave market represented the future purchase and sale of labor, the conditions of credit and the rates of interest in the local market would have a great importance in determining the price structure for slaves. Given the conditions of supply and demand in the predominant activities, it was to be expected that a structure of prices was created that rewarded physical vigor, sex, age, and skills that involve slaves of all ages, including children and the elderly. Prices reflected these supply and demand conditions and the skills of the slaves, such as the varied jobs that needed to be done, slave assignments, and the different levels of health.

It is also clear that even in Brazil there was a possibility of internal reproduction of the slave labor force. Studies have shown that in some periods and regions, there was a positive natural growth of the slave labor force. But even when this did not occur, there was a process of reproduction that amplified the participation of slaves born in Brazil and thus increased their importance in the supply of the slave labor market. The supply of slaves born in Brazil altered the general supply conditions because it created a source of slaves relatively independent of the Atlantic slave trade. It also created a labor supply completely different from the Atlantic slave trade. Instead of adults, preferably male, who were offered by the Atlantic slave trade, the internal supply offered a far more ample range of opportunities, with equality of sexes and a range of ages. This expanded the opportunities of the slave market, as much for the buyers as for the sellers of slaves. For a specialized producer needing slaves, like a miner who was uninterested in sustaining his labor force and had a slaveholding strongly concentrated in adult males, investing in the reproduction of slaves was less interesting a proposition than for a farmer dedicated to multiple crops, who may have wanted a great number of both male and female slaves equally balanced. The care of slave children, as many studies have shown, was in fact a specialized task in the larger slave units. We find that in the large slaveholdings in agriculture, there are a greater proportion of married slaves and children. Thus, as exemplified in these two extreme cases, each slave owner or potential slave owners was offered varied alternatives in terms of the age and sex composition of their slaves, depending on the work to be developed by these slaves. A slave owner could readily sell an adult ready for work for a price sufficient to acquire two young slaves, who then gradually reached maturity in physical terms, experience, and possibly of price. Others could be attended by a young child or by a woman of an advanced age, but that depended on the needs and possibilities of the buyer. These various demands and the possibility of work for the slaves permitted the elaboration of a price structure that reflected the particular demand conditions and the varied labor capabilities of the slaves.

The market for slaves was also influenced by different types of taxes imposed on slaves, from the original taxes charged by various governments in Africa or America during the transport phase to those charged in the local market. During some years, the government even charged a special head tax on slaves to their owners, as for example in the case of Minas Gerais during the early mining period. It was a way for the Crown to appropriate part of the rents being generated in the region. In some periods, the royalty taxes charged by the Crown on gold production, called the Quinto²⁹ for its value, was levied indirectly on the slaves owned, which represented the potential production capacity of each miner. Taxation was also used to limit the interprovincial slave trade after the abolition of the Atlantic slave trade. This was due to the attraction of slaves to those activities with the greatest profit, which in that period was coffee production. The only way to maintain their slave force in regions without coffee and with less profitable enterprises was through a tax on the final price of slaves in the purchasing region. This clearly prejudiced the owners of the slaves in the exporting regions, but maintained the viability of economic activity in the original slave-exporting regions. The maintenance of slave activities in the other regions was fundamental for the preservation of slavery in the third quarter of the nineteenth century when the abolitionist movement intensified. No region was denuded of slaves and all would face losses should slavery be abolished.

Numerous studies of Brazilian income and wealth in the preemancipation period have shown that slaves represented a significant part of individual wealth. Although there is variation depending on time and place, the results of these wealth studies suggest that slaves normally made up about 30 percent of the individual patrimony of wealthy persons in Brazil. Even more important was the liquidity of this share of an individual's wealth, superior in fact to the value of lands and equipment.³⁰ Besides, part of the transactions with slaves was realized via credit, which was made possible because of the great liquidity that slaves represented, especially compared to other nonfinancial assets, and for the relative stability of their price, shown in numerous studies that have analyzed slave prices. It is also important to understand the relation between land and slave ownership. Land was obtained in large part through the land grants known as sesmarias until well into the nineteenth century. These were great areas of land conceded by the Crown with the condition of effective exploitation of these grants. When one compares the average size of these land grants with the structure of slave ownership, in which small- and median-size slaveholdings predominated (few greater than 40 slaves and rarely in the hundreds), one notes an incompatibility between the size of the landed property and its potential to be economically exploited, given the size of the slave labor force employed. Even with the significant increase of the poor free population that was continually used in the Brazilian economy, its utilization was not very important in the labor force contracted for the large estates. How then could these large estates, which were minimally exploited, be maintained over generations? It would appear that between squatters and servants, most of the great property owners could defend their large estates from exploitation by allowing their lands to be settled by subsistence farmers until such time as they could begin serious commercial activity, at which time all these poor workers were expelled from the large estates and replaced by slaves.³¹

It is fundamental to understand that the extraordinary extension of slave labor in Brazil was reflected in the other forms of labor. Although during the four centuries of slavery in Brazil there occurred a great expansion of the free population, the majority of whom were free persons of color, a free, salaried labor market was never organized that would have permitted the replacement of slaves for free labor in the most important economic activities, those that were primarily directed toward the international market, like sugar and later coffee. Free labor existed, but usually in dispersed activity, not in the more complex forms of agricultural production, such as the labor gangs on the plantations or in mine production. The free worker was situated in the two pools of labor organization – either as skilled and specialized workers or those dedicated to tasks supplementary to the great exporting centers (in opening roads, clearing virgin forests, repairing fences, and so on). In addition, the persistence of slavery inhibited the use of immigrant labor, despite the vast supply of European workers in the nineteenth century. As long as slavery existed, it was not possible to effectively introduce immigrant workers into Brazil. Once slavery was eliminated, massive foreign immigration occurred in the leading coffee sectors of the Southeast, above all in São Paulo. Such was the abundant international supply of such labor that despite the abrupt emancipation of the slaves, there was no serious loss of manpower in the core Brazilian coffee fields. Immediately, black slaves were replaced by an intense influx of immigrations who effectively supplied the local labor market with workers, while elsewhere in the rural areas a complex arrangement of free labor based on the ex-slave labor force evolved that maintained local and export production in traditional crops.

How can one explain the existence of a large contingent of free persons, in large part mestiços and nonslave owners, who dedicated themselves to subsistence or lived as dependents (a favor) or servants (agregados) on the lands of others and could not be organized into a free labor market capable of substituting or even complementing slave labor? Even after the closing of the Atlantic slave trade and the increasing pressures of the international community made the ending of slavery a distinct and clear possibility, why was this substitution of slave labor impossible to achieve? From the beginning, it was patently difficult to attract free-wage workers with the characteristics of Brazil: a hostile tropical country, but more particularly one with abundant lands and a powerful oligarchy unwilling to allow open competition with free smallhold farmers. There was no system of control capable of preventing wage workers from gradually turning themselves into such small free farmers (either with lands officially purchased or as squatters) without some type of restriction on their freedom of labor. In the classic phrase of Edward Gibbon Wakefield:

What was the sole cause of the revival of slavery by Christians, but the discovery of waste [empty] countries, and the disproportion which has ever since existed to those countries between the demand and supply of labour? And what is it that increases the number of slaves to Christian masters, but the increase of Christian capitalists wanting labourers, by the spreading of Christian people over regions heretofore waste?³²

As the modern economist Evsey Domar noted, of the three elements – “free land, free peasants, and nonworking landowners – any two elements but never all three can exist simultaneously. The combination to be found in reality will depend on the behavior of political factors – government measures.…”³³ Clearly, in the Brazilian context with abundant free lands and a landowning class promoted by the Crown, the turn toward forced labor was inevitable. One has only to recall that late into the nineteenth century, the maps of the province of São Paulo showed that half the territory of the province was still classified as “lands occupied by Indian savages.”³⁴

What determined if a land was free or occupied was in truth its viability for economic exploitation, not a simple land title. This viability was in large part determined by the transport system. As long as the system of long-distance transport was exclusively based on mules, it was impossible to grow products that could be exported to the distant coast. This factor, for example, limited the potential expansion of sugar. It did not limit the extraction of gold and diamonds because of their high relative prices. However, it did limit the expansion of bulk products such as coffee, which only expanded into the interior of São Paulo via the railroad. Thus, there was always free land in abundance, either on the frontier or in previously exploited and abandoned areas – a common feature of Brazilian agriculture. The effective economic occupation of Brazil did not go beyond a few hundred kilometers from the coast until the second half of the nineteenth century. This factor explains the abundance of free nonslaveholding workers in subsistence farming and the lack of a large free labor wage market in Brazil. This was only created with the abolition of slavery and the massive immigration of free European workers at the end of the nineteenth century. It is also worth noting that the very existence of slavery, as an alternative labor force itself, inhibited the development of a free labor market. This created a prejudice against manual labor. The integration of free Brazilian workers into the labor market was a slow process that extended into the second half of the twentieth century. There are well-known cases of attempts to integrate slaves and immigrants in coffee production. But this led to strikes and conflicts within the fazendas and in turn resulted in several European countries prohibiting labor recruitment of workers for Brazilian fazendas.

Based on the very extensive studies that deal with slavery in the various regions and during different time periods, what can be said about the general organization of that labor system in Brazil? We can say that it reproduced the essential characteristics found in other regions, but there are also unique features to the Brazilian development of African slavery in America that should be emphasized. These features resulted from the special historical process of the evolution of this implanted labor system in the evolving structure of Brazilian society and economy. In turn, this labor system lasted for such a long a time that it fundamentally influenced the structure of the society and economy of Brazil until the second half of the nineteenth century.

African slave labor was initially introduced in the northeastern regions, essentially in the production of sugar, which was gradually extended throughout the sixteenth and seventeenth centuries. Probably in this period, such labor was limited to the areas influenced by sugar production, with little utilization in other areas of the colony. Other regions could marginally utilize the slaves of African origin, but essentially used native Brazilians who were available everywhere. Through forms of cooptation or compulsion, they became the basic labor force in many regions. The Paulistas, for example, true pioneers of the Brazilian backlands, were primarily based on Indian labor in all of their many and varied activities, but generally integrated them as forced laborers. By the third century of colonization, only a limited part of the present Brazilian territory was economically exploited. Settlements were spread along the coast, with little integration between them, because there was little economic reason to do so. The only part of the colony that was well integrated were the sugar-exporting regions, which in turn were based primarily on African slave labor after 1600. Although Indians had been utilized in sugar in the colony's early stages of development, they were soon replaced by a steady supply of Africans brought across the Atlantic. Indian laborers were then used outside the sugar zones to support the export activities of the sugar plantations, zones that as yet were not heavily based on the more expensive African laborers.

The discovery of gold and diamonds in Minas Gerais, Goiás, and Mato Grosso at the beginning of the eighteenth century profoundly altered the structure of territorial settlement, with colonists and their slaves and dependent workers moving westward into the interior and southward along the coast and consequently expanding Portuguese settlement. With the new mining regions at the center, the colony became far more economically integrated, with the development of a mule-producing industry in the far south and with foodstuffs and meat animals being raised in the Paulista highlands and Rio de Janeiro to supply the growing population of the interior. Roads were opened to the three mining provinces, and the mule trains that became the common form of transport on these roads led to supporting settlements along the routes to furnish the interior markets and to provide food and support for the muleteers and their mule trains. In the eighteenth century, African slave labor was being introduced in all the regions, no matter what their economic activity. São Paulo, whose labor force was once almost totally based on Indian labor, rapidly adopted African slave labor. In other areas, this probably occurred at the same time, although perhaps in a less intense way. All the recent studies show that no matter what the region of Brazil or what its local economy produced – whether exporting sugar, coffee, precious metals, jerked beef, or foodstuff production for the local market; fishing; local craft production; or professional service activities – African slaves were found to be an important presence in all areas. In the case of export crops, they represented the labor force base. In activities directed toward the internal market, in many cases they worked alongside family laborers or as assistants and journeymen for artisans and professions. They were also found in all aspects of domestic labor.

In each region, the predominant local economy determined the demand for slaves and the appropriate local price structure, whether directed toward the external or internal markets. This constant demand generated a supply of Africans slaves, not only from Africa, but from other regions of the colony as well. Although there should have been a relatively homogeneous price structure throughout Brazil determined by African supply conditions and by the slave trade, the peculiarities of the local market, particularly those some distance from the coast, initially created local variations in these prices.

This homogeneity in the national market was also affected by the geographic mobility of the masters and their slaves. As in all such expanding frontier American slave societies, slave owners were able to constantly move themselves and their slaves to newly opened economic regions. Brazil was well-known for a moving economic and settlement frontier. Exhausting the soil or looking for new economic opportunities, farmers, ranchers, and planters moved to virgin areas, temporarily occupied them, and often moved on again to another virgin area, and this frontier remained an open one well beyond the end of slavery. Labor followed these frontiersmen, and the migrating masters brought their slaves with them. There was no static labor market that inhibited the movement of slaves to these new regions or even to new types of work. Slaves followed their masters and new slaves could be bought in local or even distant markets.

Local economic conditions also had an impact on the demography of the slave population, at least in the initial periods. Traditionally, it has been thought that the activities of greatest economic intensity, usually related to exportations and the plantation model, led to an extreme exploitation of labor, leading to a low life expectancy and an inability of the slaves to reproduce themselves through natural growth. Recent studies have shown that the determining factor in the capacity of slaves to naturally reproduce is the demographic structure of the slave population itself, primarily in terms of their age and sex distribution. In the major export crop regions, there was a high demand for adult male slaves, which led to a distorted age and sex structure. A population heavily weighted toward adult males was not ideal for self-reproduction. The latest research has shown that in those regions where these export activities were less intense or declining, or in areas little influenced by them, as in Minas Gerais in the nineteenth century or São Paulo in the period prior to the rise of sugar and coffee production or in Paraná before the impact of coffee when small-scale farming and ranching predominated, natural reproduction became possible because of the better balance of age and sex of the resident slaves, in turn determined by different demand and supply conditions.

There was a generalized use of slaves in all regions and all activities from the beginning of the eighteenth century until the middle of the nineteenth century. Only after 1850, with the end of the slave trade, were slaves once again primarily concentrated in the zones of greatest economic development. During this century and a half – the period of the most intense activity of the Atlantic slave trade – slaves were well distributed throughout the colony and empire and were utilized by their masters in the most varied economic activities. It is undeniable that the masters, who utilized their slaves in these multiple activities, had a clear notion of the economic viability of their businesses and of the potential alternatives to such slave labor. To acquire a slave represented a purchase of anticipated labor, requiring the purchaser to determine the maximum price he or she could pay based on the future value of work produced, taking into account factors such as the cost of capital, the expenses for the maintenance of the slaves, expectations on the life of the slave and his or her long-term productive capacity, as well as the conditions of the market for the goods produced with slave labor. The conditions of the local market for slaves were amply known and, in function of the potential profit of each business, became viable through the acquisition of slaves most adapted to the type of activity being developed. The market supplied men and women of all ages, physical conditions, and learned or native aptitudes. Given the liquidity of this market, the farmer, artisan, or person offering services had a gamut of alternatives to exchange, sell, rent, or buy new slaves. There are even cases of sales or inheritances that involved only a share of ownership of an individual slave. In activities of low profit, they probably utilized slaves of lesser price, such as children, older persons, or adults with limited work capacity. Slaves were also extensively used in domestic labor in practically all of the activities exercised in this area. The majority of such domestic slaves were women, but women were also used in most other activities, including agriculture.

Rather unique to Brazilian slavery was the large number of slaves who hired themselves out – the so-called escravos de ganho. Although masters who rented their slaves to others for all types of labor existed in Brazil as in most American slave societies, the escravos de ganho were relatively unique to Brazil and some of the other Latin American slave societies. Slaves rented by masters were simply given to an institution, an entrepreneur, an artisan, a farmer, or other private person and received from the renter a fixed amount, usually paid monthly, with the renter typically paying for the maintenance and well-being of the slave. In this arrangement, the slave had no autonomy whatsoever, and this type of rental arrangement was as common in Brazil as in all other slave societies.³⁵ In contrast to this system, the self-renting slaves, or escravos de ganho, were given total autonomy to carry out business arrangements as long as they themselves, not some third party, paid their rental fee to their master. The essential difference, then, is that such self-rental slaves made legal contracts, worked on their own, and paid their masters a fixed fee as a daily or weekly “rental” and often paid for their own maintenance if they lived outside the home of their master. Finally, many of them, especially those who worked in the streets, were licensed by the city councils. They were thus recognized to have their own peculium independent of their masters once their fixed rentals and whatever maintenance costs they had to undertake were paid. Such escravos de ganho could live at the home of the master – and thus would have maintenance paid for them by the master, or they could live outside their master's home and then could have either arrangement – either paying their own maintenance or having the master pay it for them. If they were living in a workshop (taller) of a master craftsmen, either a journeymen (oficial) or master (mestrer), then that artisan who was training them paid maintenance, and if they worked in a factory, the same usually would have occurred with the factory owner. The maintenance of a slave has been variously estimated. One master of urban slaves in Salvador de Bahia in 1839 estimated that maintenance costs for ten slaves were $640 réis per week per slave, a rate of 91 réis per slave per day. Another estimated that his household domestic slave cost him more than double that rate, or 200 réis per day.³⁶ Assuming the higher number, it would seem that a slave who paid for his or her own maintenance would have to make at least 600 réis per day to break even, if his rental payment might be as high as 400 réis per day and his maintenance could go as high as 200 réis per diem. Some commentators suggest that after the rental fee and the maintenance costs were paid (if the slave paid them), the self-rental slave had relatively little left over in savings. But because this group often purchased their freedom, it can be assumed that some savings, however small, were an essential part of the system; otherwise there would have been no incentives for the slave to work under this arrangement. At the same time, these were the classic slave occupations for which the masters had to offer positive incentives in order to obtain good economic results.

In the court city of Rio de Janeiro between 1851 and 1879, some 2,868 slaves (only 45 of whom were women) were granted license to work in the streets as escravos do ganho na rua, which, though the most common activity, was not the only one for such slaves. Although the majority (or 95 percent) were simply described as street workers, the largest single-defined category for men was fishmongers, and for women, products of the sea, with bread and cake sellers an important second group.³⁷ In terms of origin, 2,195 (or 77 percent) of these escravos de ganho na rua were Africans.³⁸ These licenses seem to have been the norm in other centers. In Salvador de Bahia, such a registration law was passed by the municipal council in 1835, and the one for Rio was issued in 1838.³⁹

These slaves usually paid their master a previously determined sum on a daily or weekly basis and, less frequently, on a monthly account. The majority of such street slaves were either itinerant peddlers of products or porters who carried cargo either alone or in groups, on carts, on hand-pulled wagons, on their shoulders, or even on their heads in baskets.⁴⁰ Most elite persons moved through the city on litters typically carried by slave porters who were escravos de ganho. These numerous slave vendors and porters were a common sight in all the cities of Brazil and appear in innumerable paintings by Europeans depicting typical daily scenes in colonial and imperial Brazilian cities. But they were also such self-rental slaves who did not ply their trade on the streets, which included fishermen, sailors, skilled or semi-skilled artisans, and even greengrocers working in stores. They were barbers and barber surgeons and stevedores formally registered with the Customs (Alfândega) and exclusively used to unload ships. One of the more common practices was for owners to send their slaves to learn industrial skills and then, after a period of apprenticeship, to allow them to practice their new trades in shops of free persons or even to become assembly-line workers in industries. In this case, the industrialists or artisans paid the slaves a wage less maintenance costs, a part of which was then given to their owners. Such self-rental slaves even practiced begging and prostitution, again paying their owners a fixed daily or weekly income.⁴¹

There exist a few postmortem inventories that indicate how financially important these rental incomes from their slaves were. One from Salvador in 1847 noted that the average escravos de ganho (who were litter bearers – carregador de cadeira) were paying this master 400 milréis per day and had been doing so for 934 days, to date accumulating 373$600 réis gross income, which was a significant 83 to 62 percent of the price of a male slave in good health in the prime age group. These same slaves were evaluated at 450$000 to 600$000 depending on age, health, and skills. Thus, it took 1,500 days for the slave to pay back his original 600$000 purchase price. This same owner had four such adult litter bearers, plus one who was a journeyman shoemaker who also paid the standard 400 milréis per day. He had two other male litter bearers, still young (moços), who only paid 320 milréis per day, and Raquel, the only female self-rental slave, who was a washerwoman who starched clothes and paid him 240 milréis per day. All of these eight escravos de ganho lived outside the home of the master and paid their own maintenance. He also had one regular household domestic, Francisca Nagô, whose maintenance cost him 200 réis per day – his only slave maintenance costs. Because he showed no other major property, these ganho slaves (plus the young domestic slave) represented almost a third of the total value of his estate.⁴² What is interesting is that his eight slaves paid him an impressive sum of 2:689$920 réis in 934 days – which was more than half the current total value of his slaves and 17 percent of his total patrimony.⁴³

Aside from data available from postmortem inventories, such escravos de ganho were also listed for sale in the daily newspapers. Thus, in 1846 a master in Rio de Janeiro offered to sell his skilled stonemason “who gains 1$000 réis daily,” which was obviously at the high end of the salary scale because of his training and skills,⁴⁴ although another listed a boy (moleuqe) who was a skilled artisan, craft not given, and claimed he produced a daily rent of 1$000.⁴⁵ A skilled slave leather belt/strap maker was listed for sale who earned 800 réis per day, and a slave farm worker (preto de roça) was offered for sale who made 480 réis per day.⁴⁶ Surprisingly, a “robust preto” with no specific skills listed was claimed to make 640 réis per day,⁴⁷ whereas a healthy middle-aged preto de ganho, again with no skills listed, was said to make 480 réis per day.⁴⁸

There is little data for poorer workers listing ganho incomes, but the rental by owners of such workers was fairly common and from these monthly rental fees we can get a rough idea of what they would have earned if they were escravos de ganho. Wet nurses (amas de leite) were at the high end of the scale and probably could have rented for 645 réis per diem if the 20$000 rental fee per month asked for by one owner was the norm.⁴⁹ Children and adult domestics who were rented out by their owners as household workers and pages earned on average between 322 to 419 réis per diem, given the average 10$000 to 13$000 réis per month rental fees.⁵⁰

Those slaves who lived outside of their master's home often lived in tenements (cortiços), and these existed in abundance in all the major cities. In 1868, for example, in the city of Rio de Janeiro there were an estimated 642 such tenements or rooming houses for the poor, with a total of 9,671 rooms housing some 21,029 persons in ten of the eleven parishes of the city.⁵¹ Cheap housing for the poor was common everywhere and many of the criminal records show that there was little discrimination against slaves in the popular bars of the city. Thus, such escravos de ganho found a whole range of residential and leisure options open to them. It is thus not strange, although it would seem to be quite unusual, for slaves to even own shops and rent other slaves. This is the unusual case of Henrique, an escravo de ganho who owned his own small “angu” (a store that sold a traditional African drink made of corn) and who himself had enough funds left over after his paying his daily or weekly sum to his master to be able to rent the services of the slave Mariana from another master and use her as an assistant in his food shop. It turned out that Mariana was also his lover.⁵²

It is evident from all sources that urban slavery declined after the end of the Atlantic slave trade in 1850, and more and more of the traditional slave occupations were replaced by free persons. Such self-employed slaves survived in many occupations but tended to move toward the lower end of the skill scale. Thus, slaves remained important as porters and litter bearers in Rio de Janeiro until the abolition of slavery,⁵³ but the higher end skill of coachmen seemed to have been taken over by free persons before the end of slavery.⁵⁴

This ample and varied use of slaves in the rural and urban areas and throughout most of the settled areas created various forms of autonomy, control, stimulation, penalties, and relationship of slaves with their owners and with other free persons as well with other slaves. In some activities, as already noted, the intensity of labor was imposed through tight control and physical violence. In other activities, without eliminating either control or the potential for violence, incentives and prizes, material and immaterial, were crucial. In domestic labor, services, crafts, and small family farms, perhaps positive incentives predominated. In the plantation form of slave labor, physical force and structured work predominated; perhaps violence and other negative incentives were more important. These later relations of work and coexistence generated atypical situations in the Brazilian slave society. Significant ratios of manumission, many bought by the slaves themselves, amplified the free population of African origins, which in turn grew not only from manumissions but from a generalized process of miscegenation that occurred between free persons and between free persons and slaves. The form in which slavery was organized in Brazil permitted ample socialization of the slave population and the formation of families, some of which were maintained for generations. This is a subject that has become a major area of new research. Marriages, families, and kinship are themes that have recently proliferated in the research from the most varied regions of the country. These studies demonstrate that it was usual for these slave families to produce part of their food needs, working principally on Sundays, the day normally dedicated to rest. Any surplus of these products could be sold in the local market, representing a potential income for the slave. It represented an opportunity for the slave and at the same time reduced the maintenance costs of the slave owners.

The structure of the distribution of the slaves also reflects the economic and social conditions that we have discussed. There are differences, but usually the slaves were present in 20 to 30 percent of the households in the period under discussion. Slaves represented a similar percentage of the total population. Among masters, slave ownership was well distributed. Great were the number of small slave owners and few were the great planters with large numbers of slaves. Masters who owned five or fewer slaves usually represented half of the slave owners and controlled about a quarter of the slaves. In the plantation export sector, we do find planters with fifty or even several hundred slaves. And these slaves in larger holdings were an expressive percentage of the local slave population. In the opposite case of a predominance of small slaveholders, those holding forty or more slaves were relatively insignificant in terms of the total slave labor force. The average slaveholding fluctuated between five and eight slaves in most regions and rarely passed this value. This of course was in the period prior to 1850. After the closing of the slave trade, there was a tendency for slaves to be concentrated in the activities of highest profit, especially in the coffee regions where slaves were held in large concentration, usually in coffee estates with more than one hundred slaves.

In comparison with most of the major slave societies in the Americas, Brazil prior to 1850 looked more like the United States than any Caribbean society. It had roughly the same ratio of slave owners among the free population and the same ratio of slaves to total population. But even within this framework, the average slaveholdings in this period in Brazil were smaller than in the United States, and until 1850, it had fewer owners of very large holdings of slaves. Whereas the United States in 1790 looked like Minas and São Paulo – all with 6 to 7 percent of the owners holding twenty or more slaves – by the 1850s, the United States was moving toward a higher ratio of such owners, reaching 11 percent of all owners by 1850 and 12 percent by 1860. In contrast, Jamaica already had 20 percent of its owners who held twenty or more slaves by 1833 (see Table 5.1).

In the eighteenth century, the general pattern, at least for Minas Gerais and most of the municípios of São Paulo, was for a predominance of owners in the five and fewer slaves category, with a quarter or more of slaves in these smaller holdings and a relatively small share of owners in the twenty or more slaves group, and with the distribution of slaves in general being equal to the ratio in the five and under category, or fewer than half of all slaves (see Table 5.2). Although few comparative data exist for the other regions, it can be assumed from later developments that these zones of slave activity were not that dissimilar in their distribution patterns from what occurred in São Paulo and Minas in the same century.⁵⁵

But this pattern not only differed across regions, it obviously differed by type of industry and was leading to increasing levels of concentration of slaves in the larger units (holdings of twenty-one and more slaves) over time. Moreover, as can be seen in Table 5.3, even controlling for

TABLE 5.1. Distribution of Owners by Size of Slaveholding for Various Regions and Nations, 1790–1860

Notes: * In the Jamaican data, slave owners with just one slave were not broken out of the 1–5 slave owner category. We have tried to recalculate Higman's breakdowns to be as close to possible to the other regions.

Sources: B. W. Higman, Slave Population and Economy in Jamaica 1807–1834 (Cambridge: Cambridge University Press, 1976), pp. 274–5; U.S. Census of 1790 and 1850; for São Paulo and Minas, Luna and Klein, Slavery and the Economy (2003).

TABLE 5.2. Distribution of Slave Owners and Slaves by Size of Slaveholding, Minas Gerais and São Paulo, 1718/1792 (districts with more than 100 slaves)

Source: Arquivo Público do Estado de São Paulo, Listas Nominativas (mapas).

industrial activity of slaves, concentration was increasing even before the end of the slave trade. More owners and more slaves in each of the major agricultural activities in the province of São Paulo increased their share in the larger units between 1804 and 1829.

TABLE 5.3. Economic Activities of Owners by the Size of Their Slaveholdings in São Paulo, 1804–1829

Source: Luna and Klein, Slavery and the Economy (2003).

This trend was evident in Minas and other regions that already showed increasing concentration at the other end. However, it also should be stressed that this trend was not uniform across all regions. Census data from the 1870s show wide variation among regions and districts, depending on local economic conditions (see Table 5.4), but in general the trends seemed to be on the rise in the centers of export agriculture. Studies of two districts in Piauí in 1875 and another in Minas Gerais in the 1860s, although based on alternative postmortem inventories, also showed rates of distribution that differed little from the mineiro districts of the eighteenth century.⁵⁶ But more common was the pattern of increasing concentration – as measured in the GINI index of distribution – as we can see in Table 5.4. In contrast to the 40s and mid-50s found in most eighteenth- and early nineteenth-century districts, those of the second half of the nineteenth century were to be found in the upper 50s and 60s.

Finally, there is little question that the late nineteenth-century coffee estates were the most concentrated slaveholdings Brazil had ever seen, on a par with the largest sugar districts in the Northeast and those of the Caribbean in the nineteenth century. The average coffee fazendas in the most advanced coffee regions contained forty-three slaves (see Graph 5.1), and these averages suggest, based on a study of one coffee district in Minas Gerais in this period, that more than half the coffee planters owned twenty-one or more slaves and that an extraordinary 90 percent of the slaves working these coffee estates were working on these larger units.⁵⁷

TABLE 5.4. Number of Slaves and Slave Owners and GINI Index of Inequality of Slave Ownership, Selected Districts by Province, Early 1870s

Source: Marcondes, “Desigualidades regionais brasileiras” (2005), tabelas 3.4, 4.4, 5.4, 6.4, 7.4, and 8.4.

At the same time that we find slaves everywhere, we find also a complex organization of foodstuff production that involved both slave owners and those who owned no slaves. Both sold products in the market. The nonslave owners and the small farmers sold the surplus of their production above subsistence in the local market. This activity, well

GRAPH 5.1. Number of fazendas and average number of slaves in the major coffee districts of Brazil, 1881–1888 (no slaves = 23,553).

Source: C. F. Van Delden Laerne, Brazil and Java (1885), pp. 222–3.

revealed in the contemporary documents, was particularly evident in Minas Gerais, São Paulo, and Paraná, but also throughout the Northeast as well, all indicating the complexity of colonial and imperial Brazilian society and showing how important was the local and regional internal market – whose complexity and importance have only recently been understood.

¹ “Os escravos são as mãos e os pés do senhor do engenho, porque sem eles no Brasil não é possível fazer, conservar e aumentar fazenda, nem ter engenho corrente.” André João Antonil, Cultura e Opulência do Brasil, Introdução e vocabulário por A. P. Canabrava (São Paulo: Ed. Nacional, s/d), p. 159.

² “[M]edo, e somente o medo, aliás empregado com muito sistema e arte, porque o excesso obraria contra o fim que se tem em vista.” The quote is from Carlos Augusto Taunay, Manual do agricultor brasileiro (reprint of the 1839 edition; São Paulo: Companhia das Letras, 2001), p. 54. Schwartz quotes a Portuguese oberver who declared, “quem quiser tirar proveito de seus negros, há de mantê-los, fazê-los trabalhar bem e surrá-los melhor; sem isso não se consegue serviço nem vantagem alguma.” Stuart B. Schwartz, Segredos internos, engenhos e escravos na sociedade colonial (São Paulo: Companhia das Letras, 1995), pp. 122–3.

³ See, for example, Fernando Henrique Cardoso, Capitalismo e escravidão no Brasil meridional: O negro na sociedade escravocrata do Rio Grande do Sul (São Paulo: Difusão Européia do Livro, 1962); Ciro Flamarion Cardoso, Escravo ou camponês? O protocampesinato negro nas Américas (São Paulo: Brasiliense, 1987); Ciro Flamarion Cardoso et al., Escravidão e abolição no Brasil: Novas perspectivas (Rio de Janeiro: Zahar, 1988); Jacob Gorender, O escravismo colonial (3rd ed.; São Paulo: Ática, 1980).

⁴ Stefano Fenoaltea, “Slavery and Supervision in Comparative Perspective: A Model,” Journal of Economic History, 44, no. 3. (September 1984), pp. 635–68. Also using this model is Flávio Rabelo Versiani, “Escravidão ‘suave’ no Brasil: Gilberto Freyre tinha razão?” Revista de Economia Política, 27, no. 2 (Abril–Junho 2007).

⁵ “…chefe de um pequeno reinado, no qual, por governar despoticamente, e acumular as atribuições de legislador, magistrado, comandante, juiz e algumas vezes de verdugo,” Taunay, Manual do agricultor brasileiro, pp. 48–9.

⁶ Francisco Vidal Luna, “Economia e Sociedade em Minas Gerais (Período Colonial),” Revista do Instituto de Estudos Brasileiros, vol. 24 (1982), pp. 43–6. This pay arrangement was also organized for itinerant slave diamond miners in Mato Grosso in the early nineteenth century. Lucia Helena Gaeta Aleixo, “Mato Grosso: Trabalho escravo e trabalho livre (1850–1880)” (Dissertação de mestrado: PUC-SP, 1984), pp. 32–3.

⁷ Robert W. Fogel and Stanley L. Engerman, Time on the Cross: The Economics of American Negro Slavery (Boston: Norton, 1974).

⁸ Paul A. David et al., Reckoning with Slavery: A Critical Study in the Quantitative History of American Negro Slavery (New York: Oxford University Press, 1976).

⁹ Gavin Wright, The Political Economy of the Cotton South (New York: Norton, 1978), chapter 3, pp. 43–89.

¹⁰ Among the 75,000 slaves listed in the population census of São Paulo in 1829, at a time when sugar was becoming a predominant activity in the Oeste Paulista and coffee in the Val do Paraíba, 81 percent of the slaves belonged to masters who were agriculturalists, although only half resided in households that produced coffee or sugar. Thus, only 40 percent of these slaves in 1829 were involved in export agriculture. Most probably, the greater part of the slave labor force was not directly involved in export production, some being outside the market (infants and the aged) and others involved in everything from domestic service to subsistence agriculture. Luna and Klein, Slavery and the Economy of São Paulo, 1750–1850 (2003), chapter 1.

¹¹ See Barbara Weinstein, “Slavery, Citizenship, and National Identity in Brazil and the United States South,” in Don Doyle and Marco Antonio Pamplona, eds., Nationalism in the New World (Athens: University of Georgia Press, 2006), pp. 248–71.

¹² See, for example, Cardoso, Capitalismo e escravidão no Brasil meridional; Octavio Ianni, As metamorphoses do escravo: Apogeu e crise da escravatura no Brasil meridional, no. 7 (São Paulo: Difusão Européia do Livro, 1962); and Caio Prado, Jr., Formação do Brasil contemporâneo (3rd ed.; São Paulo: Editora Brasiliense, 1948). A more sophisticated re-restatement of aspects of this view is found in Gorender, O escravismo colonial, and A escravidão reabilitada (São Paulo: Ática, 1990).

¹³ Alfred H. Conrad and John R. Meyer, “The Economics of Slavery in the Ante Bellum South,” The Journal of Political Economy, 66, no. 2 (April 1958), pp. 95–130.

¹⁴ Pedro Carvalho de Mello, “Aspectos Econômicos da Organização do Trabalho da Economia Cafeeira do Rio de Janeiro, 1858–88,” Revista Brasileira de Economia, 32, no. 1 (Jan./Mar. 1978), pp. 19–68; also see Pedro Carvalho de Mello and Robert W. Slenes, “Análise Econômica da Escravidão no Brasil,” in P. Neuhaus, ed., Economia Brasileira: Uma Visão Histórica (Rio de Janeiro: Campus, 1980); and “Estimativa da longevidade de escravos no Brasil na segunda metade do século XIX,” Estudos Econômicos, 13, no. 1 (Jan.–Abr. 1983), pp. 151–79; and his Chicago doctoral thesis.

¹⁵ Flávio Versiani, “Brazilian Slavery: Toward an Economic Analysis,” Revista Brasileira de Economia, 48, no. 4 (Dezembro 1994), pp. 463–78.

¹⁶ Antonio Barros de Castro, “Escravos e senhores nos engenhos do Brasil. Um estudo sobre os trabalhos do açúcar e a política econômica dos senhores” (Tese de doutorado; São Paulo: UNICAMP, 1976).

¹⁷ Iraci del Nero da Costa, “Repensando o modelo interpretativo de Caio Prado Jr.” São Paulo, NEHD-FEA/USP, 1995, 45 p., (Cadernos NEHD, n. 3).

¹⁸ Leonardo M. Monasterio, “FHC errou? A economia da escravidão no Brasil meridional,” História e Economia Revista Interdisciplinar, 1, no. 1 (2° semestre 2005), pp. 13–28.

¹⁹ In this category, see the manual by Taunay, Manual do agricultor brasileiro.

²⁰ Marquese is one of the authors who has written interesting studies in this new line of research in Brazil. See Rafael de Bivar Marquese, Feitores do corpo, missionários da mente (São Paulo: Cia. das Letras, 2004), and his study Administração & Escravidão. Idéias sobre a gestão da agricultura escravista brasileira (São Paulo: Hucitec, 1999).

²¹ See, for example, Douglas Cole Libby, Trabalho escravo e capital estrangeiro no Brasil: O caso de Morro Velho (Belo Horizonte: Itatiaia, 1984).

²² For a good discussion of the compatibility of free and slave labor in industry, see Sérgio de Oliveira Birchal, “O mercado de trabalho mineiro no século XIX” (Belo Horizonte: Ibmec, Working Paper no. 12, 2007); Douglas Cole Libby, “Proto-Industrialisation in a Slave Society: The Case of Minas Gerais,” Journal of Latin American Studies, 23, no. 1 (February 1991), pp. 1–35; and Mário Danieli Neto, “Escravidão e Indústria: Um estudo sobre a Fábrica de Ferro São João de Ipanema – Sorocaba (SP) – 1765–1895”(Tese de doutorado: Economia/UNICAMP, 2006).

²³ Fogel and Engerman, Time on the Cross, pp. 73–5.

²⁴ Laird W. Bergad, Slavery and the Demographic and Economic History of Minas Gerais, Brazil, 1720–1888 (Cambridge: Cambridge University Press, 1999).

²⁵ Pedro Carvalho de Mello, A economia da escravidão nas fazendas de café: 1850–1888 (Rio de Janeiro: PNPE/ANPEC, 1984).

²⁶ Katia M. de Queirós Mattoso, Herbert S. Klein, and Stanley L. Engerman, “Trends and Patterns in the Prices of Manumitted Slaves: Bahia, 1819–1888,” Slavery & Abolition, 7, no. 1 (May 1986), pp. 59–67; reprinted in João José Reis, ed., Escravidão e invenção da liberdade: Estudos sobre o negro no Brasil (São Paulo: Editora Brasiliense, 1988), pp. 60–72.

²⁷ Flávio Rabelo Versiani and José Raimundo Oliveira Vergolino, “Preços de Escravos em Pernambuco no Século XIX” (Texto para Discussão no. 252; Universidade de Brasília, Departamento de Economia, Brasília, Outubro de 2002).

²⁸ See Renato Leite Marcondes and José Flávio Motta, “Duas fontes documentais para o estudo dos preços dos escravos no Vale do Paraíba paulista,” Revista Brasileira de História, 21, no. 42 (2001), p. 495–512; Maria Luíza Marcílio et al., “Considerações sobre o preço do escravo no período imperial: Uma análise quantitativa (baseada nos registro de escritura de compra e venda de escravos na Bahia),” Anais de História, no. 5 (1973), pp. 179–94; Luiz Paulo Ferreira Nogueról, “Sabará e Porto Alegre na formação do mercado nacional no século XIX” (Tese de doutorado, Campinas: UNICAMP, Economia, 2003), chapter 4; Carlos A. M. Lima, “Escravos Artesãos: Preço e família (Rio de Janeiro, 1789–1839),” Estudos Econômicos, 30, no. 3 (Julho–Setembro 2000), pp. 447–84; Afonso de Alencastro Graça Filho, “As flutuações dos preços e as fazendas escravistas de São João del Rei no século XIX,” IX Seminário sobre a Economia Mineira, pp. 147–78.

²⁹ The Crown taxed a fifth of gold produced, which was called the Quinto.

³⁰ See Table 4.5 and Chapter 4, note 44.

³¹ For an analysis of this moving frontier and its impact on class structure and slave use, see Alida C. Metcalf, Family and Frontier in Colonial Brazil: Santana de Parnaíba, 1580–1822 (Berkeley: University of California Press, 1992).

³² Edward Gibbon Wakefield, “A Letter from Sydney” [1829], in Lloyd Pritchard, ed., The Collected Works of Edward Gibbon Wakefield (Glasgow: Colllins, 1968), pp. 112–13.

³³ Evsey D. Domar, “The Causes of Slavery or Serfdom: A Hypothesis,” Journal of Economic History, 30, no. 1 (March 1970), p. 21.

³⁴ See for example the 1868 map of São Paulo produced by Candido Mendes in his Atlas do Imperio do Brasil and available at http://www.novomilenio.inf.br/santos/mapa23g.htm.

³⁵ For example, a survey of advertisements related to slaves in the Rio de Janeiro newspaper Jornal do Comercio in the month of January 1848 showed that rented slaves were offered almost every day in the newspapers, and these slaves were everything from skilled craftsmen to cooks, ironers, washerwomen, and even wet nurses who were rented out directly by their masters. Typical of such ads was that which appeared on January 18, 1846, p. 4, which offered “to rent a preta slave who knows how to wash, starch, and iron and cook for 12$000 réis per month – advanced payment (“Aluga-se uma preta que sabe lavar, engomar e cocihnar por 12$ adelantados.” On January 6, 1847, p. 4, in the same newspaper, a slave wet nurse who had given birth just two months before to her first child was offered for 20$000 per month (“aluga-se uma preta ama de leite do primero parto, parida ha dois mezes por 20$ mensaid pagos adelantados”).

³⁶ Maria José de Souza Andrade, A mão de obra escrava em Salvador 1811–1860 (São Paulo: Corrupio, 1988), pp. 134–5.

³⁷ Luiz Carlos Soares, O ‘Povo de Cam’na capital do Brasil: A escravidão urbana no Rio de Janeiro do século XIX, (Rio de Janeiro: Faperj, Letras, 2007), tabela XLVIII, p. 421.

³⁸ The single-largest group, numbering more than a thousand, came from West Central Africa; half that number came from West Africa; and 294 from East Africa. There were 347 whose African nation was not listed. Soares, O ‘Povo de Cam’na capital do Brasil, p. 128.

³⁹ Such a license was also required of free workers who sold their services or goods on the street. For these two laws, see Paulo Cruz Terra, “Tudo que transporta e carrega é negro? Carregadores, cocheiros e carroceiros no Rio de Janeiro (1824–1870)” (Diss. de mestrado: UFF, Niteroi, 2007), pp. 34–5.

⁴⁰ One of the few porter occupations exclusively identified with slaves was gasnho com cesto – that is, carrying baskets filled with all types of goods on their heads. Terra, “Tudo que transporta,” p. 40.

⁴¹ One of the best detailed studies of these ganho slaves, also known as ganhadores, is found in Soares, O ‘Povo de Cam’na capital do Brasil, chapter 5. Professional prostitutes were mostly free-colored Creoles or Portuguese women in the first half of the nineteenth century in Rio and they were eventually replaced by women from central and eastern Europe. Slave women practiced more part-time and hidden forms of prostitution, although masters did put their women to street solicitations, even after curfew hours, and arranged with them to work as escravos de ganho to encourage their activity. Luiz Carlos Soares, Rameiras, Ilhoas, Polacas…A prostituição no Rio de Janeiro do século XIX (São Paulo: Editora Ática, 1992), especially chapter V. A set of cases on slave prostitution in the 1870s suggests that poor owners, often women, put up their slaves for prostitution, but usually in secondary brothels, as street solicitation was too dangerous for slaves. In fact, it would seem that slaves doing street solicitation were escravos de ganho. Sandra Lauderdale Graham, “Slavery's Impasse: Slave Prostitutes, Small-Time Mistresses, and the Brazilian Law of 1871,” Comparative Studies in Society and History, 33, no. 4 (October 1991), pp. 670, 673.

⁴² On the relative prices of skilled and unskilled slaves by age and sex, see Lima, “Escravos Artesãoes,” tabelas 4 and 5, and graficos 1–5, pp. 478–9, 482–4. What is interesting to note is that prices for skilled workers remained at the same high plateau, from 20 to 40 years of age, whereas those for unskilled workers of both sexes dropped significantly after peaking in their 20s.

⁴³ Andrade, A mão de obra escrava em Salvador 1811–1860, pp. 133–4.

⁴⁴ “Vende-se…um pedeiro que ganha 1$000 diarios…” Jornal do Comercio, January 18, 1846, p. 4; and January 12, 1846, p. 3, for the leather workers.

⁴⁵ “Vende-se…um moleque perfeito oficial, que ganha 1$000 diarios…” Jornal do Comercio, January 7, 1846, p. 4.

⁴⁶ “Vende-se…um preto de nação, oficial de correeiro, que da de jornal 800 rs, e su precio de 600$ rs.” Jornal do Comercio, January 12, 1846, p. 4. “Vende-se…um preto de roça, que ao ganho da 480…” Jornal do Comercio, January 7, 1846, p. 4.

⁴⁷ “Vende-se por muita crecisão um reforçado preto, que ganho 640 réis por dia…” Jornal do Comercio, February 6, 1846, p. 3.

⁴⁸ “Vende-se…um preto de media idade, que ao ganho da 480 rs.” Jornal do Comercio, February 9, 1846, p. 3.

⁴⁹ “Aluga-se uma preta, ama de leite do primero parto, parida ha dous mezes por 20$ mensaid pagos adelantados.” Jornal do Comercio, January 6, 1846, p. 4. For an analysis of numerous other advertisements of rentals and sales of wet nurses in Rio de Janeiro in 1848 in the Jornal do Comercio, see the unpublished essay by Bárbara Canedo Ruiz Martins, “Meninas e mulheres: As imagens das amas-de-leite no mercado de trabalho doméstico urbano do Rio de Janeiro (1830–1888),” and for later periods, see her thesis, “Amas-de-leite e mercado de trabalho feminino: Descortinando práticas e sujeitos (Rio de Janeiro, 1830–1890)” (Dissertação de mestrado: Universidade Federal do Rio de Janeiro–UFRJ, 2006).

⁵⁰ See, for example, rental offers for five pretas at 12$ and 13$ milreis and another for a preta “que sabe lavar, engomar e cozinhar por 12$,” Jornal do Comercio, January 18, 1846, p. 4.

⁵¹ Marilene Rosa Nogueira da Silva, Negro na rua, a nova face da escravidão (São Paulo: Editora Hucitec, 1988), tabela 13, p. 126.

⁵² Soares, O ‘Povo de Cam’na capital do Brasil, pp. 140–1.

⁵³ Soares, O ‘Povo de Cam’na capital do Brasil, p. 128.

⁵⁴ On free labor in some subspecialities, such as coachmen for horsedrawn carriages, see Ana Maria da Silva Moura, Cocheiros e carroceiros, homens livres no Rio de senhores e escravos (São Paulo: Editora Hucitec, 1988).

⁵⁵ At one extreme would be the ranching communities of Paraná, which in 1818 were those owning twenty or more slaves and holding only 17 percent of the slave labor force, and the average slaveholding had just under 5 slaves (for 483 slave owners and 2,345 slaves). Horácio Gutiérrez, “Terras e gado no Paraná tradicional” (Tese de doutorado: USP, 1996), tabela 4, p. 65.

⁵⁶ Renato Leite Marcondes and Miridan Britto Knox Falci, “Escravidão e reprodução no Piauí: Oeiras e Teresina (1875)” (Texto para Discussão, Série Economia, TD-E/26, USP-FEAC-Ribeirão Preto, 2001), tabelas 4 and 5; both these districts had GINIs in the mid-50s. One of the highest GINIs recorded (.076) was for Bananal in São Paulo in 1873; José Flávio, Motta, Nelson Hideiki Nozoe, and Iraci del Nero da Costa, “Às Vésperas da Abolição: Um Estudo sobre a Estrutura da Posse de Escravos em São Cristóvão (RJ), 1870,” Estudios Econômicos, 34, no. 1 (Jan.–Mar. 2004), tabela 12, p. 191. For the Minas study, see Deborah Oliveira Martins dos Reis, “Araxá, 1816–1888: Posse de escravos, atividades produtivas, riqueza,” XIV Encontro Nacional de Estudos Populacionais, ABEP (2004), tabela 4, panels 1866–68.

⁵⁷ A separate study of just the coffee producers in the Minas Gerais município de Santo Antônio do Paraibuna from 1851–1870 based on postmortem wills also showed an average of 43 slaves per fazendas (108 owners and 4,683 slaves). Some 90 percent of the slaves resided on estates that had twenty-one or more slaves and 53 percent of the masters owned these estates. The average for these more than twenty slave estates was a very high seventy-four slaves per unit. Mônica Ribeiro de Oliveira, “Cafeicultura mineira: Formação e consolidação 1809–1870,” IX Seminário sobre a Economia Mineira, Quadro 4, p. 70.