The most reliable way for revenue authorities to find out how to improve tax compliance is to conduct large scale research experiments. To do this, they need the services of experienced researchers. The first such research experiment on the African continent was initiated in Rwanda in 2014. It is a joint enterprise by the Rwanda Revenue Authority (RRA), the ICTD and the African Tax Administration Forum (ATAF). The Commissioner General of the RRA was very impressed by this first experience of work- ing with the ICTD, and he offered a Memorandum of Understanding to encourage further research and capacity building activities. 


Discovering Which Messages Work Best 

The research team conducted an experiment to determine which types of messages are most effective for increasing compliance among payers of personal and corporate income taxes. The RRA sent messages, either by mail, email, or SMS, to 13,000 taxpayers. Some received a “deterrence” message, emphasising the sanctions and penalties for non-compliance. Others received a “benefits” message, illustrating how their tax payments are used by the government to pay for public services (the graph used in the letters is on the right). The third group received a generic “control” message to remind them of deadlines for paying their tax. This research is on- going. However, the process of conducting the experiment has already led to important organisational developments within the RRA. 


IMG 2626

The research team, comprised of members from RRA, ATAF, and ICTD 


Revelations about Compliance and Actual Progressivity

The research’s communications with so many taxpayers generated useful reactions. Two recipients got in direct touch with the top management of the RRA to explore the reasons for apparent irregularities in their tax accounts. One of these two cases has triggered a criminal investigation into a tax accountant who may have misreported several clients’ accounts. As a result, the top management of the RRA has come to appreciate the importance of effective communication with taxpayers as one way of increasing compliance.

The analysis of some of the data collected has revealed important facts about the tax system of Rwanda. On the basis of data on tax returns and audits, the researchers have estimated compliance gaps and audit rates. They discovered that small firms outside of the capital city, Kigali, are extremely unlikely to be audited. And these are exactly the taxpayers who are least compliant. It also became clear that the tax system is much less progressive in practice than it appears in principle. Smaller taxpayers bear a relatively high tax burden as a share of their income, compared to larger ones. This is because many choose to remain in the “lump-sum regime” that is much simpler to comply with and greatly reduces their chances of being audited. Presented with actual evidence on these issues, that had previously been topics for speculation, the RRA management has become very interested in the policy implications.


Improving Capacity and Internal Procedures

The research experiment has also triggered positive changes in the procedures the RRA uses to communicate with taxpayers. Having seen the benefits of automatic personalisation and mail merge, the RRA has dispensed with hand-addressed letters. The study focused attention on the value of continuously updating the contact information in the taxpayer registry, and the importance of encouraging all departments to cooperate in this. For example, more efforts are being made to ensure that auditors feed new contact information they receive into the central registry. The call centre is also considering adopting a new protocol to confirm contact information with each person calling in by phone. The research also catalysed the development of an SMS platform, which will allow the RRA to reach a larger number of taxpayers and to address each by name rather than through a general message. 




The RRA Research Department staff have also benefited from collaboration with the ICTD and ATAF researchers. Through working to implement the experiment, they have learned new research skills. They have also developed complementary research skills from a series of interactive ICTD workshops on experimental design, proposal- writing, and the use of statistical software.