Consumers in Africa increasingly pay for their purchases through mobile money, especially since the pandemic. These transactions are known as digital merchant payments. Rwandan consumers can choose between using standard mobile money services or a specific service only for digital merchant payments – MoMo Pay. Digital payments of any kind have the potential to improve tax compliance, because they imply digital data trails and better record keeping. How far is this potential being realised in Rwanda? In collaboration with the Rwanda Revenue Authority, we collected survey data from 1,100 merchants country-wide and were able to correlate this with tax administrative data, i.e. the tax records of the interviewees held by the revenue authority. We also conducted focus group discussions with 15 merchants. We found that the great majority of payments are still made in cash. Larger, more knowledgeable and financially included merchants opt for MoMo Pay as opposed to standard mobile money, the latter being preferred by female and less educated and equipped merchants. At the start of the pandemic, in March 2020, for a period of 18 months, all fees on MoMo Pay transactions were waived to foster digital payments through the service. In September 2021, fees were then reintroduced. The waiver led to a significant rise in the use of MoMo Pay relative to cash. When the MoMo Pay fee was reintroduced, there was a significant shift back to cash from both MoMo Pay and standard mobile money services, even if the latter were not affected by the fee. Lastly, we measure whether the adoption of digital payments correlates with merchants’ tax perceptions and compliance behaviour. First, we show that merchants using MoMo Pay tend to disagree with the obligation of paying taxes in order to receive public services, a measure of fiscal reciprocity. Such negative correlation is probably due to the fee imposed on MoMo Pay. Furthermore, standard mobile money usage improves the perceived ease of complying with taxes, while that is not the case for MoMo Pay. Again, the fact that fees on MoMo Pay are not clearly identifiable in MoMo Pay statements complicates merchants’ reporting and reconciliation of their activity for tax purposes. When it comes to compliance behaviour with VAT, the adoption of digital payments by merchants only improves their reported VAT sales and inputs, and only in the short term, while final VAT liability does not change. This hints at perverse compensating strategies to avoid taxes. We recommend that the tax administration better understand the adoption patterns of digital payments and incentivise usage among less equipped categories of taxpayers. The tax administration would also benefit from getting access to mobile money data to better monitor and enforce merchants’ compliance.

Many low-income countries are increasingly digitising various tax services, usually motivated by efforts to increase efficiency and transparency and reduce the burden of compliance for taxpayers. However, where awareness and adoption are suboptimal, tax e services may produce only partial benefits. In this paper, we examine the adoption of tax e-services in Rwanda, a low-income country which has invested significant resources in digitalising government service delivery and made tax e-services mandatory from 2015. Using a combination of panel survey and tax administrative data, we study the drivers and impacts of e-services awareness and adoption. We find evidence that, before the pandemic, female and less educated taxpayers, with less sophisticated businesses, were left behind in technology adoption, even where e-services were the only option for taxpayers. Exploiting the outbreak of COVID-19 during our data collection, we also study shifts precipitated by a shock that normalised digital transactions. Take-up of e-services is remarkable two years after the pandemic, but still not universal. For those not using the e services the same challenges in access persist – indicating the potential for more targeted policy interventions. Interestingly, technology adoption is not strongly related with filing behaviour, and we study the reasons why non-filers report using the tools and, on the contrary, why active filers report they do not. Also, we do not find any significant impact of e-services adoption on perceived fairness of the tax system and overall willingness to pay, which we hypothesised benefit from e-services. Finally, using evidence from qualitative interviews, we highlight practical challenges in using e-services, such as connectivity problems and slow systems, which undermine the potential benefits.

No abstract

This study has been prepared within the UNU-WIDER project SOUTHMOD -simulating tax and benefit policies for development Phase 2, which is part of the Domestic Revenue Mobilization programme. The programme is financed through specific contributions by the Norwegian Agency for Development Cooperation (Norad).

Information technology (IT) has great potential to help increase taxpayer compliance and revenue collection. Despite the increasing use of IT solutions by African tax administrations, evidence on its effectiveness remains limited. In Rwanda, the Revenue Authority introduced a more advanced version of its electronic billing machines (EBM) to enhance its ability to track business transactions remotely and to improve taxpayers’ experience of using the machines. Using a wealth of administrative data collected by the Revenue Authority, this paper evaluates the impact of the adoption of EBM2 on the ways in which firms file their tax returns. In particular, we are able to compare first-time users of EBM2, who are mostly new taxpayers, with ‘shifters’, who moved from the old EBM1 to EBM2. We looked first at value added tax (VAT). Overall, the adoption of EBM2 resulted in significant increases in reported business turnover, non-taxable sales, taxable sales, VAT inputs and VAT due. There was also a reduction in the proportion of completed VAT returns that implied zero VAT liabilities. Unsurprisingly, there was no significant overall change in the VAT returns from ‘shifters’. They had probably internalised the benefits of electronic billing machines when using the earlier EBM1 version. The effects of the adoption of EBM2 on income tax returns are less positive. Overall, no increase in income tax liability is reported. These results suggest that taxpayers do not believe that the Revenue Authority will attempt to reconcile their (separate) VAT and income tax returns. Taxpayers probably provide more reliable VAT returns because they believe, on the basis of the installation of electronic billing machines, with upgrades, that the Revenue Authority is focusing more on VAT. The main policy implication is that the Revenue Authority should make more effort to reconcile firms’ separate VAT and income tax returns, so that the positive effects of the new electronic billing machines on VAT compliance will spillover into income tax compliance.