Abstract: Abstract: In the process of state building in new territories, rulers face important trade-offs. They can create their own administration (direct rule) or delegate power to existing political institutions (indirect rule). This choice has important consequences for long-run state building and local political development. Pre-existing empirical work relies on context-dependent country-level episodes which makes it difficult to understand the determinants and consequences of indirect rule. This paper leverages a novel panel data set covering the histories of 456 chiefs and 508 episodes of village governance by armed groups in 106 villages in the DRC. We find that armed groups are more likely to co-opt chiefs when chiefs have more local authority, as measured by their coethnicity. They also rely on indirect rule when the armed group lacks legitimacy among the population, as also measured by their coethnicity. The use of direct rule increases with an armed groups' tenure and the resources of the village. We use survey data and implicit association tests to estimate the effects of indirect rule. We show that indirect rule decreases legitimacy of chiefs. Armed groups, however, increase their legitimacy by delegating power to the chiefs.
Abstract: Building a successful state requires efficient tax collection. Yet, many developing countries find themselves in a kleptocratic state equilibrium where state administrators take advantage of their positions to extract resources, citizens are reluctant to pay taxes and ultimately the state lacks resources to provide public goods and pay its agents. We provide a model of kleptocratic tax collection where the abusive power of state agents depends on their office, their connections, and superior information. We experimentally change the balance of power between state agents and the population in two ways. First, we organize pro-bono tax consulting. Second, we connect respondents with advocacy resources. To overcome difficulties in measuring taxes and bribes, we develop a smartphone application and train 300 households and businesses in the Democratic Republic of the Congo to report all payments weekly. Our results show that empowering citizens can allow them to circumvent the kleptocratic state equilibrium. The tax consulting and protection treatments empower respondents, especially unregistered firms, to engage with the tax collectors. Respondents report paying less bribes and a lengthier negotiation with state agents. Moreover, the treatment induced more interactions with state officials and increased the amount of formal taxes paid. Our findings then suggest that empowerment of citizens may be a promising avenue to move away from a kleptocratic state equilibrium.
The Lasting Impact of Colonial Educational Policies in Nigeria: Evidence from a Policy Experiment on Missionary Activity. (with Horacio Larreguy)
Abstract: This paper exploits a policy experiment on missionary educational activity in colonial Nigeria to estimate the lasting impact of colonialism on human capital, and the channels that explain its persistence. Results indicate that the treatment has a positive and large long-run effect on religion conversion, educational outcomes, health provision, consumption of durable goods, housing characteristics and occupation. The lasting impact of the policy experiment is explained by the persistence in the private provision of education which, as a result of the policy, substituted for missing public provision of education. The gap between treated and non treated areas is closing over time due to the reversion of the lack of public provision of education.
Abstract: While a large recent literature has focused on building state capacity from scratch, less is known about the ability of federal governments to induce local governments to increase their state capacity. We examine the effects of a Mexican federal program designed to incentivize municipal governments to support sustained and inclusive municipal economic and environmental development by certifying state capacity and public goods provision across 39 indicators. Certification is self-assessed and corroborated by local corruptible institutions, and may thus create incentives for political manipulation. Our difference-in-differences estimates indicate that entry into the program quickly produced large positive effects on actual state capacity outcomes. Nevertheless, we also find evidence of political manipulation: while municipal indicators showing low initial levels of performance experienced large increases in "green" certifications, they also experienced the smallest increases in actual state capacity on the variables supposed that should have informed the certifications. These negative estimates are driven by municipal governments politically aligned with the state governments - which are largely responsible for the implementation of the program - and more competitive electoral races.