Improving Tax Collection Capacity in Developing Countries

Researcher: David Lagakos (UC San Diego) and Anders Jensen (Harvard)
Location: Ghana

The goal of the project is to identify the key constraints on local government tax collection capacity, which remains low in most developing countries. This low level of tax collection leads to a low level of public goods provision. After creating a database of taxation capacity in Ghana, researchers will conduct a randomized control trial to assess the impact of introducing a combination of new technologies and new tax collection methods. 

Researchers will study the barriers and drivers of tax enforcement technology adoption through a large-scale field experiment across tax administrations in the 216 local districts of Ghana. Local tax performance in these districts is characterized by overall low levels of revenue collection,  but with a very large amount of variation in the cross-section. To study the extent to which such cross-sectional variation can be causally explained by differences in administrative capacity, the project will proceed in two steps.

First, through a novel survey methodology, researchers will establish measures of local tax administration performance and constraints. The survey tool seeks to quantify the local administration's practices in operations, targets, monitoring, and incentives. These management practices will be assessed in the context of tax registration, collection, and enforcement.

Second, researchers will implement a field experiment to study the barriers to enforcement technology adoption, focsuing on LUPMIS, an information system which enhances observability and enforcement of local property taxes. LUPMIS was developed by the Ghanan Ministry of Local Government and made availabe to all districts, but the take-up rate has been remarkably low.

After having established the clear benefits of LUPMIS adoption, the experiment will test competing hypotheses to explain lack of adoption, such as aversion to change, loss of flexibility, decreased scope for corruption, and lack of local (executive) leadership. Since both the management practices of local tax authorities and the enforcement and third-party properties of LUPMIS are similar in many regards to those in national administrations, this project's evidence will bear externally valid insights on the large variation across countries in adoption rate of enforcement technologies.

Results

Design of the survey instrument is underway.