The evaluation tested two theories of how governments can increase tax revenues: 1) governments can bargain with citizens, providing public goods or other desired policies in return for compliance with taxation; 2) governments can invest in bureaucratic capacity and monitoring to decrease the costs of collection. We find encouraging evidence that the interventions increased tax compliance, especially in the bottom-up group. At endline, vendors in the bottom-up treatment markets were more likely to have a tax payment receipt, which is the measure of tax compliance least vulnerable to measurement error and therefore the most trustworthy. While we also find higher revenues in the treatment markets at endline, we cannot eliminate the possibility that the markets in the treatment groups had different rates of revenue collection at baseline, in which case the interventions did not in fact increase revenues. Thus, while we are confident that the bottom-up treatment increased tax compliance, we are not sure whether this additional revenue actually reached the government. In the top-down group, we find no effect on individual compliance. While it is possible that we did increase the revenues that actually reach the government, the uncertainty surrounding revenue data quality means that we cannot say for certain. Finally, markets that received both the bottom-up and the top-down treatments did not experience significant increases in either tax compliance or revenue collection, which could be because the two treatments cancelled each other out or because the complicated nature of the interventions made it challenging to implement both, especially with proper attention to how the vendors might receive different components.
