Using peer information for taxation: A randomized field experiment in Ethiopia

Policy Context

Information asymmetries are a core obstacle in the collection of public revenue generation, particularly in developing countries that lack administrative capacity. In developing economies owing more than half of their employment to small and medium enterprises (SMEs), like Ethiopia, public authorities struggle to estimate SMEs’ taxable amounts correctly – a significant missed opportunity for revenue generation. In 2017, Ethiopian officials were met with protests after attempting to collect taxes based on SMEs’ average daily revenue estimates.  Currently, the Ethiopian government calculates taxable amounts by collecting information from businesses.

Can shopkeepers provide tax information about neighboring businesses? The research team will conduct a randomized controlled trial (RCT) to rigorously test and analyze the effectiveness of soliciting peer information using a tested, incentive-compatible, and low-cost intervention.

Study Design

In collaboration with Ethiopian Policy Studies Institute, Ethiopian Revenue and Customs Authority (ERCA), and Addis Ababa University, the research team will test the incentive-compatible peer information/prediction. The researchers will gather approximately 1,800 entrepreneurs, randomized by 400 blocks/clusters in Addis Ababa. The business owners (agents) will be privately asked their beliefs regarding their peers’ incomes. Then, a ranking system will be used to form a probability distribution for each agent. Agents will be compensated for rankings they provide that have empirical frequencies higher than predicted by others when indicating probability distributions. The team will randomly assign both the degree of financial incentive for accurate reporting and whether the information will be shared with ERCA. They will evaluate outcomes including reported revenue, profit, and tax payments.

Results & Policy Lessons

Results forthcoming.