How to Best Nudge Taxpayers? The Impact of a Tailored Letter Experiment in Eswatini

Abstract

Very little is known about why taxpayers in sub-Saharan Africa (SSA) remit their taxes. In collaboration with the Eswatini Revenue Authority (SRA), this study implements a nationwide randomised controlled trial nudging more than 20,000 income tax payers with behaviourally-informed mailings. This study attempts to shed new light on the drivers of SSA taxpayers’ compliance, and how can these be leveraged by resource-constrained tax authorities. While the tax nudge literature has boomed in OECD countries and Latin America, only a handful of studies can be found on SSA – this paper contributes significantly to these. First, thanks to the wealth of administrative data available, this study is the first of its kind to target three different categories of taxpayers at the same time – non-filers, nil-filers and active; most of the existing literature focusses on positive filers. Second, we tailor the content of letters to be specific to each taxpayer category. Third, we are able to target both companies and individuals, and explore heterogeneity of results along a number of dimensions, including past filing behaviour. We find that non-filers significantly respond to the nudges, while nil- and active filers do not. The best performing nudges build on the deterrence and taxpayer-assistance paradigms. Perverse responses are found from large companies. With the causal evidence produced, we are able to formulate policy recommendations on how best to target the complex ecosystem of income tax payers.