Working Papers

Local Elites as State Capacity: How City Chiefs Use Local Information to Increase Tax Compliance in the D.R. Congo

with Pablo Balan, Gabriel Tourek, and Jonathan Weigel [CEPR Working Paper] [Pre-Analysis Plan] [AEA Registration]

Forthcoming at the American Economic Review

Summary for a broader audience: VoxDev

Abstract: This paper investigates the tradeoffs between local elites and state agents as tax collectors in low-capacity states. We study a randomized policy experiment assigning neighborhoods of a large Congolese city to property tax collection by city chiefs or state agents. Chief collection raised tax compliance by 3.2 percentage points, increasing revenue by 44%. Chiefs collected more bribes but did not undermine tax morale or trust in government. Results from a hybrid treatment arm in which state agents consulted with chiefs before collection suggest that chief collectors achieved higher compliance by using local information to more efficiently target households with high payment propensities, rather than by being more effective at persuading households to pay conditional on having visited them.

The State Capacity Ceiling on Tax Rates: Evidence from Randomized Tax Abatements in the DRC

with Gabriel Tourek and Jonathan Weigel [CEPR Working Paper] [Pre-Analysis Plan] [AEA Registration]

Summary for a broader audience: JPAL, World Bank Development Impact

Abstract: This paper investigates how tax rates and tax enforcement jointly impact fiscal capacity in low-income countries. We study a policy experiment in the D.R. Congo that randomly assigned 38,028 property owners to the status quo tax rate or to a rate reduction. This variation in tax liabilities reveals that the status quo rate lies above the revenue-maximizing tax rate (RMTR). Reducing rates by about one-third would maximize government revenue by increasing tax compliance. We exploit two sources of variation in enforcement— randomized enforcement letters and random assignment of tax collectors — and find that the RMTR increases with enforcement. Replacing tax collectors in the bottom quartile of enforcement capacity by average collectors would raise the RMTR by 42%. Tax rates and enforcement are thus complementary levers. The government would raise 26% more revenue by jointly optimizing tax rates and enforcement than by optimizing them independently. These findings provide experimental evidence that low government enforcement capacity sets a binding ceiling on the revenue-maximizing tax rate in some developing countries, and thereby demonstrates the value of increasing tax rates in tandem with enforcement to expand fiscal capacity.

Optimal Assignment of Bureaucrats: Evidence from Randomly Assigned Tax Collectors in the DRC

with Pedro Bessone, John Kabeya Kabeya, Gabriel Tourek, and Jonathan Weigel [AEA Registration]

Summary for a broader audience: VoxDev

Abstract: The assignment of workers to tasks and teams is a key margin of firm productivity and a potential source of state effectiveness. This paper investigates whether a low-capacity state can increase its tax revenue through the optimal assignment of its tax collectors. We study the two-stage random assignment of property tax collectors (i) into teams and (ii) to neighborhoods in a large Congolese city. The optimal assignment involves positive assortative matching on both dimensions: high (low) ability collectors should be paired together, and high (low) ability teams should be paired with high (low) payment propensity households. Positive assortative matching stems from complementarities in collector-to-collector and collector-to-household match types. We provide evidence that these complementarities reflect high-ability collectors exerting greater effort for high payment propensity households when matched with other high-ability collectors. Implementing the optimal assignment would increase tax compliance by an estimated 36% relative to the status quo (random) assignment. By contrast, the government would need to replace 62% of low-ability collectors with high-ability ones or increase collectors’ performance wages by 69% to achieve a similar increase under the status quo assignment.

Religion and the Scope of Morality: Evidence from Exposure to Missions in the D.R. Congo

Abstract: I study how historical exposure to Christianity, a religion that emphasizes universal morality, affects individuals' attitudes towards in-group members (i.e., family and coethnics) and out-group members (i.e., non-coethnics). I use archival records to identify the location of colonial Christian missions in the Democratic Republic of Congo and combine it with contemporary data collected from surveys, social network questionnaires, and a lab-in-the-field job referral experiment. Examining a sample of 1,019 respondents, I find that historical exposure to Christian missions results in less in-group bias today. This effect is explained by more favorable attitudes towards out-group members, while attitudes towards in-group members are left unchanged. Additionally, missions are associated with more out-group members in individuals' social networks, a higher propensity to recommend out-group members for a job in the referral experiment, and more universalist self-reported values. To address the potential bias introduced by the endogenous location of missions, I use exposure to missions that were abandoned early on and for idiosyncratic reasons as a placebo test. Exposure to abandoned missions does not appear to have the same effect. Finally, I find evidence that part of the missions' enduring effect is explained by the adoption of a Christian identity that cuts across family and ethnic ties. However, this identity also engenders a negative bias towards non-Christians.

Land Formalization in Weak States: Experimental Evidence from Urban Property Titling in the D.R. Congo

with Pablo Balan, Gabriel Tourek, and Jonathan Weigel [Pre Analysis Plan] [AEA Registration]

Summary for a broader audience: EGAP Stories of Change

Abstract: Despite the importance of property rights in institutionalist models of development, there remain important gaps in knowledge about land titling in urban settings. We study a randomized land titling program that cuts the ticket price of formal land titles by 90% as well as reducing transaction costs of acquiring a title. The program increased the demand for land titles by 44 percentage points and caused a 13.7 percentage-point increase in the acquisition of titles. Exploiting randomized subsidies, we estimate an elasticity of demand for land titles of -0.42. Examining heterogeneous take-up, we find higher demand among individuals with higher-value properties, more income, and more education. We also find that take-up is highest within the footprint of the former colonial city of Luluabourg and lowest in parts of the city still under customary authority. Finally, we study the short-run effects of urban titling. Although we find no evidence of increases in tenure security, investment, or access to credit, we show suggestive evidence that the program crowded out engagement with informal institutions within local communities. Overall, the results support the idea that the formal state and local, informal forms of governance and social insurance may function as substitutes.

Fiscal Contracts? A Six-Country Randomized Experiment on Transaction Costs, Public Services, and Taxation in Developing Countries

with Ana de La O et al. [Pre Analysis Plan] [AEA Registration]

Abstract: We present results from six randomized controlled trials jointly designed to promote formalization and tax payments in low- and middle-income countries. Each randomized intervention used in-person visits, during which citizens received information about the government benefits that come with formalization and assistance in undertaking one of three types of formalization (business registration, property regularization, and access to public services). A meta-analysis shows that the average effect of these interventions on citizens’ intent to formalize, on formalization, and on tax payment is indistinguishable from zero. However, we find substantial heterogeneity across sites in individuals’ in-tent to formalize and actual formalization, which suggests that there are both demand- and supply-side barriers to formalization. The results shed light on central questions about informality and underscore the difficulty of regularizing taxation and service provision in low- and middle-income countries.


Income Concentration in British India, 1885-1946 with Facundo Alvaredo and Guilhem Cassan, Journal of Development Economics, 2017, Volume 127 pages 459-469

The Association between Income and Life Expectancy in the United States, 2001-2014 with Raj Chetty, Micheal Stepner, Sarah Abraham, Shelby Lin, Ben Scuderi , Nick Turner and David Cutler, 2016, The Journal of the American Medical Association, Vol 315, No. 14

Work In Progress

Tax Collection and Bureaucrat Performance: Experimental Evidence from the DRC

with Gabriel Tourek, Elie Kabue Ngindu, and Jonathan Weigel [Pre Analysis Plan] [AEA Registration] [fieldwork completed]

Abstract: This project examines how working as a tax collector affects local bureaucrats' performance. We exploit random variation in whether local bureaucrats known as avenue chiefs were responsible for property tax collection (treatment), or whether agents of the tax ministry collected taxes within avenue chiefs' jurisdictions instead (control). We estimate the performance of avenue chiefs by measuring how they choose to allocate scarce benefits from a government anti-poverty program in their community, based on the distribution of lottery tickets to vulnerable populations. The resulting distribution of anti-poverty benefits enables us to measure the extent to which chiefs targeted the program toward the neediest families in the community as well as diversion, nepotism, ethnic favoritism, and taxpayer reciprocity. We find that chiefs who taxed target the distribution of lottery tickets towards lower-quality properties, commit fewer inclusion errors (leakage) and exclusion errors (under-coverage). They distribute more lottery tickets to taxpayers (taxpayer reciprocity) and are less subject to nepotism and favoritism.

The Effects of Decentralized Tax Collection on Citizen Engagement

with Gabriel Tourek, Laura Paler, and Jonathan Weigel [Pre Analysis Plan] [AEA Registration] [fieldwork completed]

Abstract: This project examines how taxation by different levels of the government affects citizen's engagement. We exploit random variation in whether local bureaucrats known as avenue chiefs were responsible for property tax collection (treatment), or whether agents of the tax ministry collected taxes within avenue chiefs' jurisdictions instead (control). We propose to examine whether taxation by different levels of government induces greater citizen engagement directed towards the level of government that collected taxes. Moreover, we probe whether decentralized taxation is accountability enhancing in light of the fact that chiefs might also engage in more ethnic taxation, which could lead to more ethnic-based collective action and deepen ethnic divisions within communities. We measure citizen engagement using a novel behavioral exercise in which citizens have an opportunity to act collectively to demand a community monitoring meeting in the context of a real-world anti-poverty program. Overall, this inquiry aims to shed light on the engagement and accountability implications of taxation at different levels of government in fragile states.

Using Machine Learning to Improve Property Tax Collection in the D.R. Congo

with Arnaud Fournier, Gabriel Tourek, and Jonathan Weigel [Pre-Analysis Plan] [AEA Registration] [fieldwork completed]

Abstract: Cities in developing countries often lack the financial capacity to finance public goods. Property taxation has been identified as a promising source of revenues for cities in the developing world since (i) it generates local tax revenues, (ii) it is a relatively efficient tax, (iii) it can be progressive and capture growth in real estate values. However, many weak governments today do not collect property taxes effectively due to the absence or incompleteness of property valuation rolls. We propose to use machine learning to construct property valuation rolls in contexts where information about property values is limited. We use a small training sample of 2,000 properties and machine learning and computer vision models to predict the values of 46,000 out-of-sample properties in Kananga, DRC. We find that even with a small training sample machine learning methods that make use of survey measures of property and neighborhood quality perform well. Computer vision methods that use visual features of property pictures have a harder time learning property values in our context.

The Political Economy of Ethnicity in the Democratic Republic of Congo

with Sara Lowes, Nathan Nunn, James Robinson, and Jonathan Weigel [Pre Analysis Plan] [AEA Registration] [fieldwork completed]

Abstract: We study the extent to which politics can shape ethnic preferences. We exploit time variation in political competition between ethnic groups caused by an episode of decentralization that increases the number of provinces from 11 to 26 in the Democratic Republic of the Congo (DRC). Decentralization changes the extent to which different ethnic groups are key players in the political arena. In the region of interest, groups that were previously unimportant politically will become key players, and groups that were political competitors previously will gain majority control of their own province. We use a series of lab-in-the-field experiments, IATs, and surveys to examine how ethnic preferences change as a result of the decentralization process.