Personal Connections in Bureaucracies: Evidence from the Spanish Empire

• In Event: Historical Political Economy

Abstract

I build a novel biographical dataset of Spanish Empire bureaucrats, comprising high-ranking colonial officials (473 audiencia ministers) and their superiors in Spain (159 councillors of the Indies) from 1708 to 1808. I show that personal connections between officers in the colonies and in Spain were key to the empire’s ruling strategy.

The empire faced a formidable governance problem. All formal authority resided with the king in Madrid, but the huge distances made direct rule impractical. Letters from Spain took up to four and eight months to reach Mexico and Lima, respectively. It is therefore not surprising that, despite the theoretical centralization, colonial officials were granted considerable autonomy. In this paper, I will show evidence that central rulers in Madrid appointed agents whom they could trust (because they had previous relationships with them) and that this practice helped them to partially overcome the agency problem.

I focus on the interaction between two institutions: The Council of the Indies and the colonial audiencias. The Council (in Madrid) drafted all legislation, it was the court of last resort, and it proposed candidates to the king for all important positions in the colonies. In summary, it was the highest authority after the king for all colonial matters. The audiencias were collegiate bodies located in the colonies. They were the highest courts of appeal in their districts and they advised the viceroy or governor on all legislation. They were important checks on executive officials and they were instrumental in maintaining royal authority.

The historical record shows that their personnel policy very explicitly minimized ties between the officer and the colony in which she served. My results show that the policy also maximized connections to Spain. I exploit plausibly exogenous turnover in the Council of the Indies (in Spain) to establish that personal connections were a major factor in promotions. I estimate that being connected to one more councillor (0.73 standard deviations) increased the promotion probability of audiencia members by 3.4 percentage points (the average probability was 20%).

I also fully map the network of connections in order to show that second-degree connections were also important, but had a smaller effect on promotions. I estimate that being indirectly connected (second-degree) to one more councillor (0.22 standard deviations) increased the promotion probability of audiencia members by 0.7 percentage points. Therefore, even though the effect of an indirect connection is way smaller than the effect of a direct connection, they explain similar proportions of the variation in promotions because indirect connections were more numerous.

Unlike previous papers on connections, I show that my results are explained by confirmed personal connections, and not by shared personal characteristics. I classify two individuals as connected if they attended the same university or college, practiced law before the same tribunals, or served as ministers in the same audiencia. For all three, I also require that they coincided in time. Therefore, I am reasonably confident that individuals coded as connected knew each other personally and interacted in person in a sustained way. Conversely, I show that shared backgrounds had no effect on promotions. Therefore, my results are likely to be driven by individuals knowing each other and not by elitism or homophily.

In addition, I show that districts with more connections raised more revenue. To identify this effect, I use variation in the Council composition. Since my strategy holds audiencia composition intact, this result is unlikely to be driven by the identity of audiencia ministers (which rules out selection or favoritism as possible explanations). Therefore, this result suggests that connected subordinates were more likely to act in their superiors' best interests and that the policy may have contributed to the empire's success. The effects are economically significant. In my preferred specification, one extra connection between a colonial audiencia and Spain (one more minister connected to a councillor) is associated with an increase of 0.6% increase in revenue. One standard deviation in connections (21.3) is associated with a 13.8% increase in revenue.

The two results together imply that the governance strategy of the Spanish Empire relied heavily on loyalty. I define loyalty as a strategy that privileges partners with whom the agent interacted in the past. In other words, councillors of Indies (principals, in Spain) were loyal to their friends by promoting them more often, and audiencia ministers (agents, in the colonies) were loyal because they performed better when their friends were in office. In other words, shaping social networks was an effective way to ensure compliance in an environment with very high monitoring costs.

Author

• Marcos Salgado, Stanford University