By Professor Abel Kinyondo, Associate Professor (Development Economics), University of Dar es Salaam
In July 2021, Tanzania introduced a levy on mobile money transactions ranging from ten shillings to ten thousand shillings depending on the amount being sent or withdrawn, presented as a means of financing rural infrastructure and projected to raise something in the order of 1.25 trillion shillings a year. It was layered on charges that already existed, namely an eighteen percent value-added tax and a ten percent excise duty applied to mobile money fees, and the combined effect was considerably sharper than the headline amounts suggested.
The scale of that combined effect is the part most often left out of the retelling. Before the levy, taxes accounted for roughly twenty-three percent of what it cost a Tanzanian to move money through the system, and once the levy took effect in July 2021 that share rose to approximately sixty percent, easing only to around fifty-six percent after the first reduction. A cost structure of that shape does not sit quietly in the background of a household budget, it changes what people decide to do.
Tanzanians responded in large numbers by simply transacting less, and the shift was rapid enough to be visible within a single quarter. Between June and September 2021, person-to-person transactions fell from about thirty million a month to eighteen million, a decline of thirty-eight percent, while cash-out transactions dropped from roughly thirty-three million a month to twenty-five million. The government responded quickly, reducing the levy by thirty percent in September 2021 and by a further forty-three percent in July 2022, which brought the maximum charge down from ten thousand shillings to four thousand, and the Finance Act 2023 then removed transfers of electronic money from the scope of the levy altogether with effect from July of that year, expressly in support of financial inclusion.
Honesty requires acknowledging that the levy was not operating in isolation during this period, since the pandemic and the global economic shocks that followed it tightened household finances across the region, and any careful reading of the transaction data has to account for both influences rather than attributing the whole decline to tax policy. The sequence nonetheless remains informative, because the sharpest movements track the levy’s introduction and its successive reductions closely enough that dismissing the connection would require more explaining than accepting it.
What gives this episode its weight is what Tanzania had built and put at risk. The proportion of Tanzanians using formal financial services rose from sixteen percent in 2009 to seventy-six percent by 2023, mobile money usage grew from sixty percent of the population in 2017 to seventy-two percent in 2023, and there were more than forty million mobile money accounts in the country by December 2022. Those gains reached people whom conventional banking had never served, particularly in rural areas where an agent is far more accessible than a branch, and they are the gains a transaction levy places under pressure first.
None of this means that mobile money should go untaxed, since Tanzania faces genuine fiscal pressure alongside a real need to broaden a narrow revenue base, and an economy with a large informal sector cannot be reached through payroll taxation alone. The question is whether a charge on the act of transacting is the right instrument for reaching it, and the answer this country arrived at, through three separate corrections in as many years, was that it is not. The levy still applies to withdrawals, which is the transaction a rural user can least avoid, so the design question has not been fully settled even now.
Tanzania has run this experiment more thoroughly than any of its neighbours and has the data to show for it. The useful contribution at this stage is to ensure the findings inform the next budget cycle rather than being quietly set aside, and that governments still expanding comparable levies have the chance to learn from evidence generated here instead of repeating the exercise at the expense of their own citizens.