On 30 September 2026, the GSMA brought mobile money taxation back into the African policy debate. Its analysis highlights a recurring problem: a levy that looks small on paper can materially change the price of a transaction, the use of mobile money accounts and the ability of the service to remain affordable for everyday payments.

The issue is not a simple choice between raising public revenue and protecting financial inclusion. The real questions are what is taxed, who ultimately bears the charge and how users respond. The statutory rate answers only part of that equation.

In Cameroon, for example, a levy equal to 0.2% of the value of electronic-money transfers and cash-outs increased the average consumer price of taxable transactions by about 19.5%, according to an IMF working paper published in December 2025. The gap between 0.2% and 19.5% is not a contradiction. The first figure is calculated against the amount transferred; the second measures the increase in the total cost relative to the service fee that existed before the tax.

That distinction is the starting point for better policy.

The statutory rate is not the economic cost

Consider a transfer whose value is high relative to the operator’s fee. A tax calculated on the amount sent can exceed the commercial fee even when the tax rate appears negligible.

The IMF study illustrates this mechanism using Cameroon’s tariff schedule. In some bands, transfer fees were fixed or increased more slowly than the amount sent. A proportional levy of 0.2% was therefore small as a share of the transfer value but substantial when compared with the original price of the service.

The right denominator depends on the question. The tax burden is measured against the transaction value. The user’s price impact is measured against the pre-tax fee. Distributional effects require yet another comparison, against household expenditure or income.

The same levy can look modest under the first measure and heavy under the other two.

What the evidence actually supports

The IMF paper combines two levels of analysis. The first draws on data across African markets. The second uses transaction records from one mobile money operator in Cameroon and the Central African Republic, with matched users in Mali serving as a comparison group.

Within that sample, the average monthly value of taxed transactions fell by 40% in Cameroon and 47% in the Central African Republic relative to the comparison group. Transaction counts fell by 33% and 51%, respectively. These are estimates produced by a specific empirical design. They should not be presented as a census of every operator or every user in those markets.

The cross-country analysis also finds a delayed effect. By the second year after a mobile money tax is introduced, active accounts are estimated to be about 20% lower and transaction numbers more than 10% lower than in the counterfactual scenario without the tax.

The status of each source matters. The IMF document is a working paper published to encourage debate, not an official policy position of the institution. The GSMA represents an industry and makes policy recommendations. The convergence of their findings deserves attention, but every conclusion should remain tied to its method and author.

Users do not respond in the same way

Banked customers are more likely to have an alternative such as a bank transfer, card or another formal payment channel. They can therefore respond more strongly to a price increase by switching instruments. Unbanked people and customers in underserved areas have fewer substitutes. Their use may decline less, but the levy absorbs a larger share of their expenditure.

This difference helps explain how a mobile money tax can be both inefficient and regressive. Customers who can avoid the taxed rail do so more easily. Those who depend on mobile money continue using it to a greater extent and carry a relatively heavier burden.

A shift back to cash is another risk. The IMF paper finds suggestive evidence of substitution towards cash withdrawals. The authors qualify this result, but it raises a fundamental question: a tax that is easy to collect on a digital payment may weaken the traceability and formalisation that made the payment attractive to the tax authority in the first place.

Ghana shows that a levy can be reversed

Ghana introduced a 1.5% levy on certain electronic transfers in May 2022. The rate was reduced to 1% in January 2023, and the levy was repealed on 2 April 2025 through the Electronic Transfer Levy Repeal Act.

That sequence does not prove that every payment tax should be abolished. It does show why a levy on a digital channel should be evaluated after launch using data on revenue, transaction volumes, transaction splitting, substitution between channels and the impact on low-value payments.

A review or sunset clause can make that evaluation more credible. Without one, a tax may persist even when its original assumptions no longer hold.

Five tests before taxing a mobile money transaction

A robust decision should answer five questions before implementation.

1. What is the tax base?

Taxing the amount transferred, the service fee or the provider’s income produces very different incidence. A tax on the transaction value can become large relative to the service price when fees are low.

2. Who pays after the market adjusts?

The legal taxpayer is not always the party bearing the economic cost. A provider may pass the levy through, absorb part of it, change its tariff structure or reduce investment. Agent remuneration may also come under pressure.

3. Are small payments protected?

The GSMA’s comparative research recommends considering exemptions for transactions below $10 and, depending on market conditions, up to $20. Those thresholds should not be copied mechanically. They need to be converted into local currency, adjusted for price levels and tested against transaction splitting.

4. Are comparable channels treated consistently?

Taxing mobile money differently from an equivalent bank transfer can shift activity without genuinely broadening the tax base. Greater neutrality between instruments reduces that risk, provided policymakers account for differences in customer profiles and cost structures.

5. How will the effects be measured?

The evaluation framework should specify indicators before launch: net revenue, active accounts, transaction count and value, distribution by value band, cash-out volumes, incidents, agent activity and the user’s average cost. Aggregate outcomes can conceal a very different effect on low-value transactions.

For providers, taxation becomes an operational problem

A new levy is not merely a percentage added to an app. It changes the pricing engine, ledger, receipts, refunds, reversals, pending-transaction handling and reconciliation with the tax authority.

The provider must separate principal, commercial fee and tax; retain the rule and rate used for every transaction; process failure recovery correctly; prevent double taxation when a transaction is retried; and disclose the full amount before confirmation.

Exemptions create additional controls. A per-transaction threshold can encourage splitting. A daily or monthly cumulative threshold requires a consolidated view of customer activity. An exemption based on customer category depends on reliable identification data and sound governance over status changes.

Every regulatory change also requires regression testing. The most sensitive cases include reversed transactions, partial refunds, interoperable transfers, cross-border payments and differences between initiation, settlement and reporting dates.

The right question is net yield

The debate should not stop at gross collections. Revenue must be assessed against administrative cost, possible volume declines, substitution into cash, pressure on agents and the effect on financial inclusion.

Payment digitalisation can help governments identify economic activity, make filing easier and broaden the tax base. Automatically taxing the most visible digital movement can produce the opposite result if users move to less traceable channels.

Mobile money taxation should therefore be treated as payment policy as well as revenue policy. The rate matters. But the tax base, exemptions, available substitutes, distribution of the burden and quality of operational execution determine the real outcome.

Sources

Cover credit: photograph by Minette Lontsie, [Mobile Money au Cameroun](https://commons.wikimedia.org/wiki/File:Mobile_Money_au_Cameroun.jpg), licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). Image cropped to a horizontal format.