Centralization of Justice and State Building

Although a central role of monopolizing conflict resolution is cementing power and legitimacy, mainstream theories of state building often overlook this key process, putting more emphasis on taxation or warfare.^{Footnote 9} The logic of controlling the administration of justice originates in a fundamental agency problem: under weak monitoring, state agents have strong incentives to deviate from the ruler’s agenda.^{Footnote 10} For early states that relied on private tax farming, local elites had incentives to overcharge the peasantry for their own benefit.^{Footnote 11} For modern states with a more developed bureaucratic apparatus, this simply means that individual bureaucrats might be involved in acts of corruption. A common denominator in these agency problems is that the general population is a potential partner for the ruler because the people have strong incentives to address grievances to the central powers. If justice is denied to them at the local level, the general population can be transformed into whistleblowers of bureaucratic malfeasance. From the rulers’ perspective, it’s often more efficient to address “fire alarms” than to deploy constant “police patrols” to monitor executive agents because the former involves less active intervention.^{Footnote 12}

Throughout history, rulers also have had an incentive to adjudicate people’s grievances, as it bolstered their legitimacy among the general population and exercised social control.^{Footnote 13} By positioning themselves as a source of protection against the rapacity of local elites, central rulers sought to increase their support among the masses. An additional outcome of a centralized court as a legitimate institutional channel to adjudicate grievances was that it offered an alternative to violent collective action: for the peasantry, petitions to the central powers became part of their repertoire of contention.^{Footnote 14} Ultimately, the capacity of centralized courts to control local agents, increase rulers’ legitimacy, and adjudicate grievances explains the prevalence of petition-making institutions across history and political regimes.^{Footnote 15}

This article illustrates the process of the centralization of justice administration in the context of Spanish America. As stated above, I use the example of the General Indian Court of colonial Mexico, a centralized tribunal established in 1592 to address indigenous grievances.^{Footnote 16} I argue that as in similar instances of state formation where the ruler confronts agency problems, this court was formed to create a direct link between the general population (Indians) and the ruler to restrain the ruler’s local agents. A key element of my argument is that the indigenous population’s access to justice was subject to the Spanish Crown’s strategic considerations and economic goals. In short, the role of the gic was never intended to end the exploitation of the Indians, “but to restrain excesses [of the local elites and bureaucrats]; in short, to keep exploitation within measure.”^{Footnote 17}

Agency problems in the Spanish Empire

The colonial enterprise of the Spanish empire, which started in the sixteenth century, was subject to a fundamental agency problem. Beginning with its first Caribbean settlements, the Spanish Crown was aware of unique opportunities to obtain rents through the exploitation of Indian labor, resource extraction,^{Footnote 18} and territorial expansion. Achieving those goals with limited state capacity became a critical problem for it. Throughout the colonial period, the Spanish empire relied on settler entrepreneurs to establish its presence in the New World, which effectively made those settlers state agents in need of monitoring. For this reason, the Crown faced a constant tension between its dependence on local elites and the perils of giving them ample opportunities to overexploit the indigenous population and accumulate wealth and power. Although the colonial state increased its presence and built a complex bureaucratic architecture, the Crown’s capacity to monitor provincial bureaucrats remained limited. Controlling local elites and provincial bureaucrats became a fundamental challenge.

The Spanish Crown was interested in preventing local agents from expanding their economic power beyond what the Crown allowed. In the Spanish empire, local elites were stakeholders of a colonial system that relied heavily on public-private partnerships and cooptation. These elites played a crucial role in tax collection and revenue allocation in the colonies.^{Footnote 19} During the early colonial period, such partnerships took the form of encomiendas, in which conquistadors received grants to extract resources from native labor.^{Footnote 20} During the later colonial period, local elites’ wealth also depended on the Crown’s trade monopoly and the concession of special privileges (fueros) to interest groups like miners and merchants.^{Footnote 21} This economic structure was designed to offer local autonomy to provincial elites and to position the Crown as the ultimate arbiter and regulator.^{Footnote 22}

Local elites were constantly seeking to expand their power and wealth by encroaching on the Indians’ territories and exploiting their labor. The encomendero elite used their control of Indian labor to amass fortunes and to seize control of local governments. On some occasions, encomenderos exercised their positions of power to obtain or even grant themselves portions of communal Indian lands. The paradigmatic example of land accumulation is the Marquessate of the Valley of Oaxaca, a vast seignorial estate granted in 1529 to Hernán Cortés, the commander of the conquest of the Aztec empire, and his descendants in the aftermath of the conquest.Footnote ²³ The encomenderos’ political power and economic influence were a threat to the Crown even from the first Caribbean settlements in the early sixteenth century. This problem led to the publication in 1512 of the Laws of Burgos, which regulated the colonists’ behavior with particular emphasis on their treatment of the native population.Footnote ²⁴ But when the Laws of Burgos failed to restrain the Spanish settlers’ abusive behavior, a new code, called the New Laws, was promulgated in 1542, giving the viceroys power to create a judicial system and prohibiting Indian slavery.Footnote ²⁵

During the second half of the sixteenth century, the encomienda was replaced by a debt-peonage system known as repartimiento (mita in Peru).Footnote ²⁶ Under this system, local elites increasingly maximized production by overexploiting the Indian labor force. At the same time, the Spanish owners of the large agricultural estates known as haciendas expanded their properties and increased productivity by invading Indian communal lands. Price fluctuations, storage limitations, and crop irregularities often amplified the burden of the debt-peonage system. Hacienda and ranch owners increased the labor obligations of the Indians until their production needs were met, which had catastrophic consequences for the native population.Footnote ²⁷ As expressed in one contemporary account, “These Indians, it should be noted, are assigned to the repartimiento for only one week, but the farmers, faced with the loss of their crops … keep them … two or three weeks and more … when he [the Indian] returns he finds his wife or children dead, and his fields ruined for lack of cultivation, or destroyed by cattle.”Footnote ²⁸

As part of the process to gain greater control over colonial affairs, the Crown created a large and complex bureaucratic apparatus. But this effort was tarnished by agency problems similar to those experienced with the local elite. The Spanish empire’s bureaucratic architecture relied on provincial administrators, known as alcaldes mayores, to conduct affairs—from tax collection to policing—at the local level.Footnote ²⁹ These officers also had the power to name lieutenants (tenientes) to rule over smaller jurisdictions. The officers’ extensive powers, coupled with their low pay, led to abuses of all sorts. Unlike local elites, alcaldes mayores didn’t have a direct interest in land accumulation.Footnote ³⁰ For their primary source of wealth, they used their positions to manipulate prices in the labor levy system (repartimiento) and distribution of goods (repartimiento de mercancías).Footnote ³¹ Defying royal legislation, alcaldes mayores constantly traded with subjects, charging excessive prices and issuing loans at considerable interest rates.Footnote ³²

Besides bureaucrats and landed elites, the Crown also delegated responsibilities to indigenous elites, and relied on them to maintain control of indigenous villages. Native leaders known as caciques were crucial for maintaining the social order inside communities, collecting royal tributes, and linking the Spanish and Indian worlds.Footnote ³³ Although the Indian elites’ sphere of influence was limited to the village level, the Crown still had an interest in controlling their power.Footnote ³⁴ Indigenous elites usually tried to overexploit Indian labor without proper payment and to obtain some benefits during tax collection.Footnote ³⁵³⁵ These actors used their positions not only to accumulate wealth, but also to kindle rebellion or opposition—as occurred in the late eighteenth century with the uprisings led by Túpac Amaru in Peru and by Jacinto Canek in Yucatan. The Crown’s strategic interest was to keep these actors strong enough to maintain social control in the villages and counterbalance non-Indian elites, but weak enough that they didn’t acquire much power or regional influence.

The collapse of the indigenous population in the sixteenth century was a significant concern for the Spanish Crown because it threatened the colonies’ fiscal potential. In central Mexico, the native population declined from about twenty-two million to less than one million,^{Footnote 36} mainly because of a series of epidemics that began with the first contacts with Spaniards in 1520. These epidemics originated in pathogens brought by the European settlers, for which the natives had no immunity. Referring to the one of the most devastating epidemics, the chronicler Fray Juan de Torquemada wrote in 1576, “A great mortality and pestilence that lasted for more than a year overcame the Indians … the place we know as New Spain was left almost empty.”^{Footnote 37} Abuse and mistreatment amplified the damage caused by the epidemics, and the surviving native population became malnourished, underclothed, and overworked.^{Footnote 38}

Around 1550, the Crown sought to reduce these negative impacts with a land distribution policy,^{Footnote 39} giving each indigenous town seventytwo hectares of communal arable land.^{Footnote 40} These communities, which became known as pueblos de indios, held a certain level of governance autonomy, meaning that—at least in theory—Spanish settlers couldn’t dwell there or own property.^{Footnote 41} Despite such efforts to isolate the indigenous population and protect them from the local elites’ overexploitation, the demographic collapse made it even easier for settlers to amass large areas of the communal land that had been left vacant.^{Footnote 42} As a result, the native population’s agricultural potential was much reduced, and the burden imposed on them by provincial bureaucrats only escalated. And while the indigenous population declined rapidly, Spaniards and criollos (the term used for the progeny of Europeans born in the colonies) were relatively unaffected, so provincial bureaucrats seeking stable rents imposed heavier burdens on smaller communities. For the Crown, the sharp population decline meant a drop in royal tribute and the loss of a counterbalance to settlers’ power.

Overall, the Spanish Crown’s governance problems can be summarized as follows: the Crown needed to keep the local elites and bureaucrats weak enough that they wouldn’t overexploit the indigenous population, but strong enough to maintain a productive colony. It was also in the Crown’s interest to keep the indigenous population and its leadership strong enough to maintain a productive colony and to counterbalance local elites, but weak enough that they wouldn’t successfully rebel against the colonial state. These governance problems, I argue, led to the creation of a centralized court. The lack of executive human power and the difficulties in monitoring individual bureaucrats made it clear that the Crown had to rely on a reactive strategy to control local agents and to be ready to put out fires wherever smoke was detected. By allowing and even encouraging complaints against individual bureaucrats, the judicial system made it possible to identify wrongdoers without threatening the foundation of the empire.

The General Indian Court

The General Indian Court was established in 1592 to redress abuses against Indians. Located in Mexico City, the court was headed by the viceroy, the king’s alter ego in the colony, who acted as the ultimate judge. For more than two hundred-fifty years, Indian claimants came from all over New Spain^{Footnote 43} to air their grievances directly to a group of officers who advised them, helped them write formal complaints, and served as translators.^{Footnote 44} On many occasions, the Indians received favorable rulings in this court, winning battles on such pressing issues as forced labor, tribute, and land rights.^{Footnote 45}

The petitioning process had clearly defined steps, which were internalized and mastered by indigenous claimants. An Indian complaint was launched when a community delegation or an individual arrived in Mexico City and met with a court agent (solicitador). These agents acted as figurative and literal gatekeepers to the court, since they could reject what they considered trivial matters. Next, the claimant’s representative would present the case to the viceroy and the audiencia, a government body with several magistrates holding executive and legislative authority. If necessary, the court summoned and interrogated the accused party and other witnesses. After the investigation, a specialized court bureaucrat (relator) summarized the case and submitted it to the prosecutor and to the chief adviser, who recommended a verdict. Then the case was sent to the viceroy for his ultimate approval.^{Footnote 46}

Part of the gic’s brilliance was that enforcement required very little of the Crown’s financial and human resources. First, the court’s specialized bureaucratic body was funded by a tax levied on Indian communities. Historian Woodrow Borah describes this as a system of “justice by insurance” because that tax contribution gave Indian communities access to the legal system at no additional cost.^{Footnote 47} Second, as a reactive body, the court was able to function with only a small number of elite officials who had high enforcement powers and low incentives for corruption. Crucially, the court gave these enforcers several tools, such as removing local officials, imposing fines, seizing assets, or exiling rebel subjects to the Philippines, to execute its rulings.^{Footnote 48}

A key element for the functioning of this centralized court was that its officials were relatively isolated from the local elites’ influence. This doesn’t mean corruption and bribery didn’t exist, but there wasn’t enough of it to derail the court’s functioning.^{Footnote 49} Moreover, this concern was allayed by the monitoring system and the checks and balances established in the colonial governing body overseeing the court. The viceroy shared power and responsibilities with the audiencia. Although formally, the viceroy presided over the audiencia, this institution had the jurisdiction to curb the viceroy’s abuses of power.^{Footnote 50} This jurisdictional overlap and the dual scrutiny system served to control the colonial government in favor of the Crown and facilitated the establishment of a centralized court.

For local elites, there were two reasons why the incentives to defy the Crown’s laws competed with pressures to acquiesce to gic rulings. First, as established above, their wealth and prestige depended on the Crown’s trade monopoly and concession of special privileges. In the complex economic architecture of the empire, these exemptions defined the elites’ capacity to capture rents and their right to adjudicate cases in differentiated tribunals. Second, to control large-scale indigenous revolts, the elites needed the support of the central power. The imperial deterrent to Indian rebellion was a fundamental pillar of the Spanish empire’s stability and a crucial element for elites’ cooptation.^{Footnote 51} During the colonial period, the Crown strongly signaled that it would invest all the resources needed to crush rebels for however long it took to do so.^{Footnote 52} Therefore, although local elites were constantly seeking ways to increase their wealth, they were also constrained by their dependence on the Crown. As John Coatsworth writes, throughout the colonial period, “elites endured, mostly in silence, an imperial regime that occasionally expropriated wealth by defaulting on the public debt, debasing the currency, and imposing forced loans and predatory taxes, especially in time of war.”^{Footnote 53}

The centralization of Indian justice in colonial Mexico also had broader consequences related to the relationship between the indigenous population and the Crown. To function, the colonial system required the Indians’ willing participation. Even after the huge population collapse, during the entire colonial period, the indigenous population significantly surpassed that of the European settlers.^{Footnote 54} Despite the conquerors’ military advantages, their control over the native population would have been untenable without the support of the native population at large.^{Footnote 55} Although Indians were certainly coerced into exploitative colonial practices, they legitimized the imperial project inside the colony by subscribing to its paternalistic policies and recognizing the Spanish king as the ultimate protector.^{Footnote 56}

For the Indians of New Spain, the gic transformed the Spanish king from an abstract and distant protector into a tangible institution.^{Footnote 57} At the same time, Indians became the first line of enforcement for the king’s laws and the viceroy’s orders. Like other centralized courts that functioned as a legitimizing instrument of the central power, the GIc provided a disenfranchised group with access to justice and agency to shape colonial policies and laws, thus becoming a genuine “weapon of the weak.”^{Footnote 58}

Rule by Law in the Spanish Empire

The gic exemplifies a process in which judicial institutions emerged not because there were limits on the ruler’s behavior, but because the Crown had the capacity to exercise control over local elites and the general population. This case contrasts with that of England after the Glorious Revolution, the paradigmatic case of rule of law development. According to Barry Weingast,^{Footnote 59} the Bill of Rights of 1689 resulted from an elite consensus that created self-enforcing limits to the Crown’s taxation and coercion powers. Under this perspective, the rule of law emerged after elites managed to overcome collective action and coordination problems and the Parliament became an equal partner in government. The new institutional arrangement allowed the government to commit credibly to protecting property rights and investors.^{Footnote 60}

In contrast, there were no limits set on the coercion and taxation powers of the Crown in Spanish America. Instead, the law imposed limits on the local economic and political elites as an instrument of political power.^{Footnote 61} This political manipulation of the law, along with the law’s lack of clarity and predictability, places Spanish America in a scenario that can be characterized as rule by law rather than rule of law. Interestingly, the strategic limits imposed by the gic had broader consequences for the general population because they helped to protect indigenous communities and their property rights. Although extraction was undoubtedly a central and destructive feature of the empire’s wealth building, the Spanish Crown was in some respects committed to protecting the indigenous population and addressing their demand for justice. The imperial judicial system not only allowed the indigenous population to resist blatant abuse, but it also gave them access to land independent of the colonial elite.^{Footnote 62} In short, the gic illustrates how autocratic rulers can use a limited form of the rule of law to curb predation by powerful elites vis-à-vis the rest of the population.

The Spanish Crown’s capacity to advance its interests—by controlling the economic elite and the indigenous population—can be seen in the fact that the gic’s rulings were never intended to establish limits on taxation. The Spanish empire developed a well-organized money-extraction machine designed to provide large surpluses to the metropolis.^{Footnote 63} The most significant tax for indigenous communities was the tribute tax (tributo), levied on all households in the Indian towns.^{Footnote 64} Indigenous communities repeatedly complained to the gic to obtain reductions and exemptions from this fiscal burden.^{Footnote 65} Throughout the colonial period, claims related to taxation had little chance of a favorable ruling, particularly in cases where those claims might reduce the colony’s productive potential for the Crown.^{Footnote 66}

Although the Spanish Crown had a large capacity to use the judicial system to advance its interests, it wasn’t entirely unconstrained, especially when it came to claims related to the Catholic Church. This institution, a powerful actor and ally in the colonial enterprise, received special treatment in cases related to the abuse of indigenous communities. Church members interacted daily with indigenous communities and often misused their power.^{Footnote 67} gic claims illustrate how these actors coerced, manipulated, and exploited Indians as much as local elites and bureaucrats did. But claims brought against the Church rarely received a favorable ruling in court.

The Catholic Church had played a crucial political role in Spain since early medieval times, acting as a foundation of the Crown’s sovereignty and as an intelligence service to control unruly subjects.^{Footnote 68} In the colonial period, troops of parish priests and missionary friars carried out the spiritual conquest that facilitated many of the interactions with indigenous groups. The Catholic Church’s critical position as a partner in governance and the colonization enterprise explains why the Crown was reluctant to limit the abuses of this institution’s members. This pattern provides further evidence that the gic “was not so much a court of law as a court of compromise and accommodation.”^{Footnote 69}

Mapping the gic’s Decision-Making Process

To represent more clearly the court’s strategic behavior and to delineate the theoretical expectations of my argument, I link the actors, their incentives as introduced above, and the expected outcomes of the claims. Because the gic ruled explicitly and implicitly in favor of the alleged perpetrators of abuse in a large proportion of cases, it opens the question of whether these decisions were based on legal grounds or on the ruler’s strategic agenda. To disentangle these motivations, we need to understand the court’s decision-making process. More generally, this analysis also seeks to provide some empirical evidence for the argument that the centralization of justice has served to cement power across history by deliberately weakening—or strengthening—local power holders and petitioners and to act as a balance of power.

Building on existing studies and a close reading of gic cases, I develop a probabilistic model of the criteria used by the court to make its strategic rulings. In this model, the court’s strategic decision to rule in favor of the indigenous population is a function of three dimensions:^{Footnote 70} 1) the topic of the case, 2) the actors involved, and 3) the local balance of power. Although the court would undoubtedly consider the merits of the case, I argue that strategic concerns outweigh those merits. The probabilistic model implies that not all high-merit cases will receive a favorable ruling, but that the likelihood of it will be much greater with cases that originate in a province where the Crown has strategic con- siderations.^{Footnote 71}

Most cases considered by the court were related to four main topics that had varying levels of strategic interest for the Crown.^{Footnote 72} One prevalent topic found in gic cases concerns complaints against the royal tribute. The model predicts that these claims would have a lower probability of a favorable ruling because they diminished the Crown’s fiscal goals. A second topic identified in the claims is associated with clashes within Indian communities. In general, the Crown acted as a mediator of internal conflicts and avoided siding with a particular party.^{Footnote 73} For this reason, the predictions for this topic are uncertain.

A third topic is related to land conflicts. Disputes over productive land have been a constant source of struggle in rural Mexico.^{Footnote 74} The gic cases include claims about land limits between indigenous towns and haciendas, invasions of communal land, and conflicts between indigenous towns.^{Footnote 75} As the Crown actively sought to protect indigenous lands, the theoretical expectation is that these cases would have a higher probability of receiving a favorable ruling. From a strategic point of view, the Crown was interested in limiting the accumulation of land by local elites.

A fourth topic refers to the physical mistreatment of Indians and other widespread abuses. These claims include complaints about physical violence, incarceration, and unlawful working conditions. In the model, claims related to this topic would have a higher probability of a favorable ruling because these abuses were a direct affront to the Crown’s decrees, as they diminished the Indian towns’ productive capacity and increased the potential of indigenous rebellion.^{Footnote 76}

As previously established, indigenous communities were threatened by potential extraction by local elites, bureaucrats, indigenous leaders, and Church members. Because the Crown’s primary strategic objective was to control its agents, we should expect that claims brought against local elites (European and criollo) and bureaucrats would have the highest probability of receiving a favorable ruling. Because the Indian leaders’ sphere of influence and their capacity to threaten the Crown’s interest were more limited than those of other state agents, we should expect that claims against these leaders would still receive a favorable ruling, though to a lesser degree. And because a ruling against Church members was against the Crown’s interest, the theoretical expectation is that claims against these actors would have a lower probability of a favorable ruling.

In addition to topics and actors, the court also considered the local balance of power in its decision-making process.^{Footnote 77} The first element determining the local balance of power was the vulnerability of the indigenous population. I theorize that when the indigenous population was in decline, the court was more likely to grant a favorable ruling to en sure that the colony remained productive and profitable for the Crown. At the same time, the disappearance of indigenous towns could potentially leave lands vacant and at the mercy of local elites, lessening the ability of the indigenous population to act as a counterbalance to local elite power.Footnote ⁷⁸

The second element shaping the local balance of power was the relative strength of the local elite. Their power varied according to their ability to control local resources and the productive potential of the provinces. Moreover, the expansion of the state’s local presence as a plausible counterbalance wasn’t uniform, with some provinces remaining under the administrative control of local elites until the end of the colonial period.Footnote ⁷⁹ In the model, Indian claims have a higher probability of a favorable ruling when local elites concentrate higher power levels, because these were a greater threat to the Crown.

Table 1 summarizes the theoretical expectations for how the claim topics, alleged perpetrators, and local balance of power conditions would increase or decrease the probability of a favorable ruling for indigenous claimants. Within each of these dimensions, the associated parameters indicate whether these are related to an increase or decrease in the probability of a favorable ruling. Under the assumptions developed in this section, cases related to land or abuses, cases involving local elites or bureaucratic actors, and cases originating in areas where elite power is high and the indigenous population is vulnerable would be the most likely to receive a favorable ruling. Conversely, cases about taxation, cases involving the Church, or cases originating in places where elite power is low and indigenous population strength is high would be less likely to receive a favorable ruling.

Table 1

Dimensions and parameters of Decision-Making in the gic ^a

^a The plus and minus signs refer to the associated probability of receiving a favorable ruling.

Measuring Favorable Rulings

The primary variable of interest is the probability that a document offers a resolution favoring the Indians. To create the dependent variable, I use a dictionary method identifying the presence of a set of unigrams and bigrams commonly related to a favorable ruling.^{Footnote 88} The legal language of the Spanish empire relied on a rigid structure and juridical jargon derived from the Roman codes; as such, these terms were very stable across time. Table 2 illustrates the coding scheme of the key terms I use to classify a document indicating a favorable ruling.

Table 2

Key Terms Related to protection

Although an imperfect measure, these words and phrases are clearly associated with the probability of ruling in favor of the indigenous population. On average, around 13 percent of the documents between 1592 and 1820 contain some of these terms. These documents unveil an interesting temporal variation in favorable rulings. Figure 1 shows how the average of protection-related documents was much higher during the seventeenth century (27 percent) than the eighteenth century (5 percent). There is also less variance during the eighteenth century, when the number of protection-related files seems to have reached a plateau.^{Footnote 89}

Figure 1

PROPORTION OF DOCUMENTS BY TYPE OF RULING (1590–1820)^a

^aTemporal distribution of documents with a favorable outcome within the Indians subarchive. The documents with a protective outcome are those that contain the terms shown in Table 2.

Measuring Topics and Actors

To identify the topics listed in the probabilistic model, I also rely on a rule-based method. As argued, I expect the colonial court to be more inclined to rule favorably for claims addressing land and mistreatment, because these two topics would have a negative effect on both the indigenous population and the Crown. Conversely, I expect the colonial court to be less responsive to claims that diminish its own power (such as tax claims) or claims not directly pertaining to local elites (such as conflicts within Indian communities). Table 3 shows the coding scheme for these topics, the total number of documents coded in each category, and the percentage of documents referring to a favorable ruling in each category. Concretely, for each document a topic becomes a binary indicator taking the value of 1 if the document contains one of these key terms. Note that this classification allows a document to cover several topics at the same time.Footnote ⁹⁰

Table 3

Key Terms for Topics

Although applying rule-based methods to the archivists’ annotations yields an accurate representation of topics and favorable rulings, identifying actors is much more challenging because mentions of hacienda owners, alcaldes mayores, encomenderos, or other non-Indian actors might not necessarily mean that these individuals are the alleged perpetrators. The summary of the 1684 case presented above is an excellent example. In it, the alcalde mayor is mentioned, but the accused party is the Indian governor, Martín Flores. In some cases, the presence of three or more actors further amplifies the problem.

To overcome this issue, I employ a second data set resulting from the unpublished personal notes of the historian Lesley Byrd Simpson, who classified thousands of gic cases in the 1930s and ’40s. I manually coded these notes, which are stored in the Bancroft Library at the University of Berkeley. To my knowledge, this is the first systematic classification of this data set.^{Footnote 91} An advantage of the Lesley Byrd Simpson Files (lbsf) is that they explicitly identify each case’s plaintiff and defendant, resulting in a more accurate classification of the actors involved. For example, they specify whether the perpetrator was an alcalde mayor, a local lieutenant, or a hacienda owner. To test my empirical framework, I classify the actors identified in the lbsf according to the taxonomy from Table 1. I again employ the rule-based method to identify a favorable ruling and the topic of each case.^{Footnote 92}

Measuring Local Balance of power

A first element in the local balance of power dimension is the strength of the indigenous population. Gerhard’s encyclopedia provides statistics of the indigenous population from the Relaciones Geográficas, a census-like data collection that served as the basis for calculating taxes.^{Footnote 93} Although not all provinces have information for every census, I was able to record a population count for the years 1550, 1570, 1600, 1645, 1745, 1789, and 1808 for most provinces. The data provided by Gerhard usually relate to Indian taxpayers (tributarios).^{Footnote 94} I use linear interpolation to calculate the population by decades. Following the results from Emily Sellars and Jennifer Alix-Garcia,^{Footnote 95} I assume that when the indigenous population was in decline it was more vulnerable to predation.^{Footnote 96}

To measure the relative strength of local elites in a given province, I compute the percentage of the indigenous population under the encomienda system. The central assumption for this proxy is that local elites would have a higher capacity to control the resources in a province when the entire population is under this system of indirect rule. The opposite would be the case if the Crown has absolute control of a province. Gerhard’s monographs provide information about the en- comiendas in each province, including their duration. I use the data of localities in the Suma de Visitas ^{Footnote 97} to identify all localities listed by Gerhard in a given province and to compute their percentage contribution to the province’s total population.^{Footnote 98} I then use each of the census collection periods as a cutoff and the date of dissolution to determine the percentage of a population living under encomienda during that period.^{Footnote 99} Although this method assumes that the population distribution within a province doesn’t change over time, it does weight larger localities differently, thus providing a granular representation of the transition to direct rule and local elites’ power.

Topics and Balance of Power

The theoretical framework of the court’s decision-making process predicts that the cases most likely to receive a favorable ruling are those that are related to land or abuses, that involve local elites or bureaucratic actors, and that originate in areas where elite power is high and the indigenous population vulnerable. The colonial courts data set allows us to test the interaction directly between the topics and the local balance of power dimension.

Specifically, I estimate model 1 for a claim i in province p in year t, in which the dependent variable indicates a favorable ruling for a claim in province at year.^{Footnote 100} The presence of the different topics is measured as a binary indicator. Because there are reasons to believe that the characteristics of the province and the timing of the claim might influence the decision, the model includes year and province indicators (λ and γ, respectively). To account for differential growth trajectories in the rate of a favorable ruling, the model also controls for an interaction between time and province (ϕ).^{Footnote 101} Additionally, because the sampling process that allows us to study these cases in the first place might vary by province, I cluster the standard errors at this level. Because the gic was located in Mexico City, several cases are mislabeled as having been produced in this location, thus overrepresenting this province. Moreover, the gic had jurisdiction over criminal matters in Indian communities surrounding the city, which are beyond the scope of this article. Thus, I exclude cases from this province. The results appear in Table 4.

Table 4

Favorable Ruling, Topics, and Balance of Power ^a

*p < 0.1, **p < 0.05, ***p < 0.01; all models include province and time fixed effects; standard errors are clustered at the province level

^a Linear regressions. The dependent variable is a binary indicator for favorable ruling according to the classification from the rule-based method. The explanatory variables are produced from the same coding process. Column 1 shows the effect of the claim’s topic on favorable ruling using the entire sample. Columns 2 and 3 test the effect of population vulnerability. A province is considered in population decline if the indigenous population is smaller than in the previous decade. Columns 4 and 5 show the effect of elite power. Provinces with powerful local elites are those in which the portion of the population in encomienda is higher than the period average. Columns 6 and 7 show the effect of claims’ topics after conditioning on population decline and different levels of elite strength.

These results show that unconditioned to local balance of power, land and mistreatment-related cases had a higher likelihood of favorable rulings (Table 4, column 1). To test the effect of indigenous population vulnerability, I divide the sample by cases with declining and increasing Indian populations (columns 2 and 3). I measure population change as the ratio of the indigenous population in a province-decade to the population of that province in the previous decade. A ratio larger than one represents a population increase.

Figure 2 (a) displays the predicted probabilities of a favorable ruling conditioned on the document’s topic and population change. The results of this exercise show that as anticipated by the theoretical framework, the predicted probability of a favorable ruling for land-related cases jumps from 13 to 61 percent when the population is more vulnerable. Similarly, the predicted probability for mistreatment rises from 9 to 50 percent when comparing provinces with increasing and declining populations. These results seem to be driven mainly by provinces within the empire’s core.^{Footnote 102}

Figure 2

Predicted Probability of Favorable Ruling Conditioned on Topics and Local Balance of Power^a

^aPredicted probability is calculated from estimates in models presented in Table 4. Population change is measured by decade. Provinces with strong local elites are those in which the indigenous population under encomienda is higher than the period average. Vertical lines over the point estimates represent the 95-percent confidence interval. The dashed line represents the average favorable ruling under population decline (panel (a)) and the average ruling when local elites are strong (panel (b)). The dotted line represents the average favorable ruling when population is increasing (panel (a)) and the average favorable ruling when local elites are weak (panel (b)).

An additional step to provide a full test for the balance of powers dimension requires considering the strength of the local elites (Table 4, columns 4 and 5). I classify cases as originating from a province with strong local elites whenever the percentage of the population living under encomienda is larger than the period average.^{Footnote 103} Figure 2 (b) depicts the predicted probability of a favorable ruling after computing model 1 on a subsample of cases with strong (weak) local elites.^{Footnote 104} The predicted probability of favorable ruling for land-related cases goes from thirty-one to forty-three when local elites are strong. Likewise, the predicted probability for mistreatment goes from twenty-six to thirty. But this category has a significant overlap in its predicted probabilities. The result suggests that the strength of local elites was more significant for the decision of granting favorable ruling for land cases than for mistreatment cases.

To thoroughly test the implications of the balance of power dimension, I run the model on a subsample of cases with declining populations and different levels of local elite power. The results appear in Table 4, columns 6 and 7. We observe that when the population is declining and local elites are strong, claims related to land conflicts are more likely to receive a favorable ruling than claims related to other topics. Although the coefficient for mistreatment is also positive, it is not significant. When comparing the coefficients between land-related cases with powerful local elites versus those with weak elites, we observe that the likelihood of receiving a favorable ruling is higher for cases with strong elites. By contrast, the likelihood for mistreatment cases isn’t significantly different for these two scenarios. Together, these results suggest that population decline had a strong effect on both land and mistreatment cases, whereas the strength of local elites affected only land-related cases. Because land was the primary source of enrichment for local elites, these results support the theoretical expectation of the court’s strategic behavior.

To further test the interdependence between topics, population decline, and local elite strength, I also compute a fully interactive model in Appendix 2. The results from this model, presented in Table A2.1, confirm the insights of the subsample analysis: the effect of population decline on the likelihood of a favorable ruling is positive and significant for both mistreatment- and land-related cases. At the same time, the effect of local elite power on the likelihood of a favorable ruling is positive and significant for land-related cases only.

ACTORS

As explained above, one shortcoming of the colonial courts data set is that it might not provide an accurate representation of the actors involved. For this reason, I turn to the LBSF data set, which specifically identifies plaintiff and defendant in each case. Using this data set, I run a model in which the dependent variable is the binary indicator of a favorable ruling and the explanatory variables are the different actors (see the results in Appendix 2, Table 2.1). Figure 3 displays the estimates for the entire sample and different subsamples based on the topics of the claims. As the theory predicts, bureaucrats, local elites, and Indian leaders have a positive relationship with favorable rulings, while the presence of Church actors has a null effect on favorable rulings.

Figure 3

Estimates of Favorable Ruling by Perpetrator and Type of Claim^a

^aPoint estimates and 95-percent confidence intervals for the effect of actors on favorable ruling conditioned on case topics. The results can be found in Appendix 2, Table A2.3.

The subsample analysis offers a more detailed explanation of the decision-making process. First, the effect predicted in the theoretical framework is consistent for land-related cases. That is, claims against elites and bureaucrats have a greater likelihood of receiving protection than claims involving Church members. In fact, claims in which a Church actor is a perpetrator are associated with a negative and statistically significant coefficient. Interestingly, contrary to the theoretical expectations, when the complaint is against indigenous leaders, the probability of receiving a favorable ruling is higher than when non-Indian elites are accused. This result emphasizes the Spanish Crown’s interest in restraining the accumulation of power and wealth by native leaders. But because the confidence intervals for the Indian leaders estimate overlap with zero, it’s impossible to provide a conclusive interpretation for these actors.

The estimates for taxes follow a similar pattern, consistent with the theoretical expectations of the probabilistic model. In these cases, different actors were involved in tax collection as agents of the Crown, usually overstepping their responsibilities by extorting money from taxpayers. Again, the involvement of local elites and bureaucrats in tax-related cases yields a positive correlation with a favorable ruling, whereas the involvement of Church actors is negatively correlated with the success of the claim. And the involvement of Indian leaders is correlated with a higher likelihood of a favorable ruling. In this case, the effect is strongly positive and significant, suggesting that the Crown was particularly interested in controlling indigenous leaders when they used their position to undermine the imperial fiscal apparatus.

None of the estimates for either mistreatment or community cases is statistically significant. The results for the mistreatment cases should be interpreted together with those analyzing the universe of cases (see Figure 2). As these results show, mistreatment cases in general are associated with higher levels of a favorable ruling; the fact that the involvement of individual actors doesn’t modify this likelihood suggests that the Crown cared about Indian mistreatment whoever the perpetrator was. Similarly, for community-related cases, the specific perpetrator doesn’t seem to affect the probability of a favorable ruling. But it should be noted that only 5 percent of the cases in the LBSF are associated with this topic, thus limiting the interpretability of these results.

Overall, the analysis of actors using the LBSF shows a strategic court behavior and a strong interest in controlling local elites, bureaucrats, and especially Indian leaders over claims related to land and taxes. When cases were related to mistreatment and community, the actors’ identity seems to be less relevant than other strategic considerations, such as population change.

Addressing the Selection Bias of Cases

One limitation of the results reported above is the potential problem of sample selection bias.^{Footnote 107} Scholars of judicial politics have argued that the biased nature of how cases arrive at court might affect the interpretation of the court’s behavior.^{Footnote 108} For the purposes of this article, the crucial question is how does selection bias in court cases limit the interpretation of a strategic court? An unbiased court would make rulings based only on the merits of a case, meaning that no factors are consistently associated with the likelihood of an (un)favorable ruling. In contrast, a strategic court would weigh factors other than the merits of the case (for the gic, these are the case’s topic, the actors involved, and the local balance of power). But it’s also possible that these factors determine the probability of appearing in court in the first place. If this selection bias exists, then court rulings might appear strategic when in fact they’re only a reflection of the types of cases that arrive at court.

In a centralized court like the gic, two types of selection bias can influence the characteristics of the cases observed. The first type is related to the structural characteristics of the province in which the cases originated. The second is litigant-driven, which means that the cases are a function of decisions made by the plaintiff or defendant. I analyze both biases below and explain why, given the existing data, neither is likely to threaten the interpretation of a strategic court.

STRUCTURAL BIAS

Structural bias can limit our ability to identify whether the gic’s decision is based on strategic grounds because some inherent characteristics of the provinces might be shaping the amount, type, and quality of claims. If the population is under stress, for example, there could be more meritorious cases than otherwise. To explore this bias, I first present the relationship between the inherent characteristics of a province and its total number of claims.^{Footnote 109} Figure 4 shows that a province’s specific characteristics are related to the total number of cases produced. In particular, provinces farther from Mexico City and with a multilingual population are less likely to bring claims to the court. At the same time, provinces with a greater extractive potential tend to present more claims.^{Footnote 110}

Figure 4

Determinants of Total Claims^a

^aEstimates and 95-percent confidence intervals from column 1 in Table A4.1 in Appendix 4. The dependent variable is the total number of claims in a province by decade.

The results presented in Figure 4 also bring attention to the fact that provinces with a larger indigenous population present more claims. In contrast, provinces with stronger local elites and those where the indigenous population is declining tend to present fewer claims. These results suggest that provinces in which the local balance of power doesn’t favor the indigenous population aren’t necessarily producing more claims and thus, they are not overrepresented in court. Therefore, there’s no selection bias associated with the sheer number of cases. But it’s possible that even if these provinces bring fewer claims, these claims can still have greater merit, for example, by representing more violent land invasions or harsher physical abuses. Such differences in merit may explain why these cases are more likely to be ruled favorably.

To address this issue, I develop a measure of merit by using the length of a document as determined by its number of pages. Given the geographical limitations and the implicit costs of presenting claims in Mexico City, indigenous communities often compiled several complaints in a single document with multiple chapters. Arguably, documents composed of chapters are more comparable because they point to a higher number of abuses, theoretically carrying more merit than documents referring to a single case. I focus only on cases in the highest quartile of page length.^{Footnote 111} An analysis of this subsample of more complex cases shows that these tend to be more balanced on their covariates and more comparable (see Appendix 4, Table A4.3).^{Footnote 112}

I use this measure of merit to determine whether certain provinces tend to bring more meritorious cases. Table A4.4 in Appendix 4 shows that neither demographic decline, the power of local elites, nor the topic of a case is correlated with this measure of merit. These results strongly suggest there’s no evidence of structural bias. To further test the prevalence of strategic considerations, I run model 1 conditioning on high-complexity cases only. This model considers the characteristics of the province and controls for temporal effects and crucially, for the topic of the case. Results are displayed in Table A4.5. This analysis shows that even after limiting the analysis to high-complexity cases, the results generally hold. Cases related to land and mistreatment are more likely to receive protection, whereas cases related to taxes and community issues are less likely to receive a favorable ruling. Similarly, the local balance of power has the predicted effect with land-related cases: when the population is declining and local elites are strong, cases have a greater likelihood of favorable ruling.