With its new fintech tax, Ghana will slow down the anti-poverty effects of money transfer within the country, boost inflation and stifle innovation.

The government faction in Ghana’s parliament recently passed a controversial 1.5% tax on digital financial services, known as e-levy, after a walkout by the opposition, which is still trying to block the tax in the courts.

Ghana, like Zimbabwe, now has one of the steepest and widest-ranging taxes on financial technology (fintech) services in the world. Most African countries have restricted taxes on mobile money, but Ghana and Zimbabwe levy the tax on all fintech transfers: bank to mobile wallet, mobile wallet to bank, and bank to bank.

In Nigeria, the tax is fixed at 50 Naira ($0.12) once the transaction amount exceeds 10,000 Naira ($23.98), while in Uganda, the tax is applied during withdrawal to incentivise inter-wallet transactions and reduce the producer cost-inflation effect. However, the Ghanaian and Zimbabwean approach is to apply the tax at the sending stage at a constant rate of 1.5%, regardless of the amount transferred.

Given the high tolerance for inflation in Zimbabwe’s current fiscal context, Ghana’s choice of role model is rather curious. We will return to the inflation point later. Let’s start with the broader civil society context.

Stifled policy debate

Despite widespread opposition, Ghana’s independent policy community (think tanks, research-oriented NGOs, academics, etc) has been less vocal and seemingly divided on the new tax. This is generally because the impact of tax on fintech is heavily dependent on multiple design factors.

The nature of digital financial services is such that tax can have multiple features and tweaking one or the other could have a huge impact on the effects. Independent policy analysts, such as those at IMANI, where I hold an honorary executive position, have therefore focused on debating different designs rather than taking a stance for or against the tax per see.

The absence of a unified civil society front thus means that the opposition to fintech tax failed to galvanise sufficient traction to stop its introduction. Uganda’s opposition was so fierce that the levy was first withdrawn and then slashed by 50%, while in Malawi, the plan to introduce fintech tax was totally abandoned. The same happened in Cote d’Ivoire, where the 0.5% tax was quickly withdrawn, and Benin, where the outcry was so loud that the government retreated even before the plan could be presented.

The Ghanaian government’s haste and seeming desperation to pass the tax bill quickly stemmed from a bid to reassure international investors of the viability of its fiscal consolidation plans, in the face of growing debt distress and estrangement from international capital markets.

The government’s posture has been one of disinterest in any serious engagement in critical policy debates. Its aloofness has led to a seriously flawed fintech tax design, widespread consumer disaffection and lack of critical political elite consensus.

Data naïveté

It is not too difficult to understand why the government was willing to expend so much political capital on such an unpopular tax. At first glance, as the table below suggests, digital payments – especially mobile money transactions – seem to represent a very large, neglected, tax base in the Ghanaian economy.

Bright Simons: Why is Ghana’s fintech tax not finding favour?

A naïve interpretation of the data would suggest a ‘transactional base’ of nearly $160bn.

Take for example cumulative transactions by one individual amounting to $13 a day (100 units in local currency). If transfers between accounts and wallets and digital payments for government services are omitted from this base – in accordance with the e-levy policy – and the transactional base is thus reduced to about $60bn, the original rate of 1.75% can easily be shown to net about $1.15bn a year – the amount initially projected by the government as likely revenue from the e-levy. At the currently approved 1.5% rate, that amount reduces marginally, but remains at par with major tax categories like import duties. Not a big deal, but very significant nonetheless.

Even so, this analysis does not take into account the fact that the huge transactional value is due to rapid electronic multiplication of a much smaller ‘real base’ of money in the mobile money (MoMo) segment. It is much smarter to start the whole analysis from the current ‘float’ of actual funds deposited in mobile money wallets versus funds in bank accounts in order to understand how quickly MoMo transactional values can drop if something like tax dampens sentiments.

A few charts are presented below to clarify further.

Bright Simons: Why is Ghana’s fintech tax not finding favour?

High-convenience fintech systems like MoMo can dramatically inflate the transactional effect of the actual funds stored across wallets, known as the ‘float’. Source: Bester, Hougaard & Chamberlain (2010)

Bright Simons: Why is Ghana’s fintech tax not finding favour?

MoMo transactional value inflation in Ghana is far more intense than other forms of fintech, thus bloating total e-money transactional value.

The CARE model

The high susceptibility of e-money to ‘transactional value inflation’ also suggests easy deflation if sentiment is threatened. Such deflation can quickly vaporise a huge chunk of the expected gains from tax leveraged with the overall transactional value in mind rather than just the float. For example, in Ghana, the total e-money float is about $800m (a portion of that amount reflects regulatory safeguards).

Despite being considerably lower than bank deposits (at roughly $18bn), MoMo transactional value – as seen in the table above – is nonetheless 25 times greater than electronic bank payments, despite bank deposits being more than 20 times greater than the MoMo float. Preserving transactional value therefore requires paying close attention to policies that affect consumers’ willingness to transact. Four design factors are critical in this regard: Caps, Aggregates, Rates & Exemptions (CARE).

The ‘Caps’ refers to the thresholds below and above which the tax rate does not apply (i.e. lower and upper bounds). Ghana, like Zimbabwe, has no upper bound to taxable amounts. The 1.5% tax rate applies linearly to any amount transferred through electronic means, even if it is investment capital or other high-value transfers related to business operations and economic production. The daily cumulative cap of about $13 means that anyone using any fintech platform for any activity other than basic personal expenses is expected to pay.

Unlike traditional VAT, the fintech tax does not distinguish between input and output economic activities. The aggregation model is thus naïve. It simply tallies payments, and once the $13 bound is breached, a linear rate starts to apply.

The fintech rate in Ghana and Zimbabwe is made all the more significant as it covers all types of digital payments, not just mobile money. The two countries also continue to apply taxes to the fees charged by the fintech operators, the approach preferred by countries like Kenya, Congo and Cote D’Ivoire. Commentators tend to see the two approaches – taxing the amount transferred directly or applying the tax to the transfer fees/charges – as competing alternatives, but clearly this is not the case in Ghana or Zimbabwe.

Bright Simons: Why is Ghana’s fintech tax not finding favour?

Ghana exempts the recipient from paying fintech tax, but charges the sender while in Uganda it is the other way round. Ghana also exempts payments made through the government’s own fintech platform for specified services and taxes. This concession obviously does not capture the many instances where tax is applied on a service or product paid through normal commercial channels. In those instances, citizens will pay the e-levy on top of VAT or other taxes already levied.

Quasi-government payments, such as those made in the cocoa sector by licensed private buyers, are not exempted, which disrupts emerging models in that space. Innovative services in the government-controlled cocoa sector, and elsewhere, shall be impacted by the e-levy. The e-levy policy also seeks to exempt payroll, ATM withdrawals (thus incentivising cash use), and loan repayments, but these concessions raise administration issues that I will address later.

A design mess

By now, it should be already evident that in the rush to introduce the tax, Ghana threw caution to the wind in some very important respects. But there is more to be said.

A careful look at the MoMo business model, using only the most popular fintech product to make a point, from the perspective of telecom operators (‘telcos’), as it has evolved since its introduction to Ghana in 2009, gives some important hints about the place of fintech in the digital economy.

Bright Simons: Why is Ghana’s fintech tax not finding favour?
MTN Ghana’s rate card for mobile money transactions. Source: MTN

It is evident from even a cursory glance at such business models that the telecom operators, who should be considered the domain experts in this terrain, exempt specific services from certain charges for strategic reasons. They use absolute amounts where it makes sense and percentage (ad valorem) charges where that is most suitable. They exempt fees altogether to incentivise uptake and elevate rates when they need to influence behaviour in a certain direction. The e-levy has now sent a wrecking ball through most of these market-driven strategies.

Another evolved, market-driven, ecosystem on the verge of disruption is the MoMo agent network. In Ghana, just as in many parts of Africa, many users send and receive MoMo transfers through agents. From about 180,000 in 2019, some analysts estimate the number to have grown to more than 300,000 today.

Bright Simons: Why is Ghana’s fintech tax not finding favour?
Agents are the beating heart of MoMo in Africa. Source: Ivan Ssettimba (2016)

The e-levy, as currently designed, does not accommodate the need for agents to send money on behalf of others. The omission, unfortunately, renders the $13 per day exemption meaningless for many people in rural areas since the agent will be forced to add the fee regardless of the amount the sender is transferring.

Due to a massive rural-urban and north-south divide, such agent-driven remittances have become critical to poverty alleviation in Ghana, accounting for more than 50% of total income in some rural households.

Bright Simons: Why is Ghana’s fintech tax not finding favour?
Differences in poverty rates in the historical 10 regions of Ghana. Source: Baah-Kumi & Lee (2021)
Bright Simons: Why is Ghana’s fintech tax not finding favour?
A glaring North-South Poverty Divide. Source: Sumaila Chakurah (2015)

Impact on economic production

The aforesaid design flaws can seep into some crevices of the economy. Consider the fact that the whole point of industrialisation is value addition. Value addition however requires an increase in the number of productive steps in key economic activity chains. To the extent that value chain actors usually make payments related to the production process electronically, they cannot avoid the new digital payment taxes.

The transfer taxes thus accrete in a nearly linear fashion alongside the actual value addition, but without distinguishing between inputs and outputs, leading to a snowballing effect, as crudely illustrated below.

Bright Simons: Why is Ghana’s fintech tax not finding favour?

The e-levy’s sponsors appear unaware of the likely contribution, in its current design, to produce inflation – with eventual pass-through to consumers – or they simply don’t care. Any tax with systemic effects pervasive enough to add as much as an additional 3% to inflation ought to be treated with serious caution. At the very least, its design must benefit from detailed comparative analysis and modelling.

The minimum expectation most analysts had was a Dynamic, Stochastic, General Equilibrium (DSGE) model to examine the e-levy’s potential impact on aggregate demand in the economy, and resultant downward pressure on other consumption taxes. The government refused to do any of that. In fact, none of the e-levy’s features were set on the basis of empirical research.

What is clear, nevertheless, is that in its current design, the e-levy belongs to the class of taxes that ablates economic efficiency by distorting producer prices.

Bright Simons: Why is Ghana’s fintech tax not finding favour?
Not all taxes are made equal, some are less efficient than others. Credit: Carnahan (2015)

But isn’t necessary innovation painful?

E-levy enthusiasts in Ghana have hyped the novelty of the tax and suggested at every turn that like the Communication Service Tax (CST), early disquiet will dissipate as the e-levy starts to perform. Innovation, they say, is sometimes painful, even when totally necessary, except that e-levy type fintech taxes are actually not that novel.

Apart from the fact that about 10 countries in Africa have in the last decade experimented with taxing digital financial services, none have seen massive revenues from doing so and there is a much longer history of governments seeking to tax transfers. The Ghanaian e-levy is a lot like the taxes that countries like Vietnam, Philippines and Tajikistan used to levy on remittances received by citizens from overseas. The Philippines abolished theirs in 1995.

Vietnam followed suit in 1997 and Tajikistan in 2002. After abolishing the taxes, Vietnam witnessed a growth in remittances of about five times in the ensuing decade. Recipients bought goods and services in the local market and paid lots of other consumption taxes, offsetting the foregone government income.

One can actually go as far back as the 18th century for insights into transfer taxes. In 1782, the UK introduced a ‘stamp duty’ on cheques, which means every time a person wrote a cheque to pay for anything they paid tax to the government.

Bright Simons: Why is Ghana’s fintech tax not finding favour?
A Barclays cheque from 1964 showing a stamp duty embossment.

Faced with much the same considerations, the government introduced various exemptions on the same arbitrary grounds that e-levy assumptions have been situated today. In 1791, anyone cashing the cheque within 10 kilometres of where they lived didn’t have to pay. In 1804, the range was widened to 15 kilometres.

Until in 1918, they stopped pretending that the exemptions had any basis. They then removed them and the tax rate even doubled. As late as 1956, the UK government was still making various rate and design adjustments to optimise collections. Then, finally, in 1971, it became clear that the tax was a waste of everyone’s time and then finally abolished, nearly 200 years after their introduction.

In short, the logic animating the desire to tax e-levy is very well known to history, but poor design choices remain a consistent bane.

Perennial tax policy incoherence

Ghana’s historic struggles with optimal tax design have often been as a result of very poor attempts to transparently and analytically grapple with the core logic of what the state is trying to achieve in the first place.

For example, whilst the gift tax eligibility threshold in the US is $12.06m ($24.12m for couples), it is $6.5 in Ghana. Essentially, almost every gift is taxable. Consequently, almost everyone liable for gift tax in the US pays it, but almost no one similarly eligible in Ghana does.

When the country introduced VAT in 1995 to replace the Sales & Service Tax, the purpose was to track value more effectively in the economy. However, since then, the country has vacillated between VAT that tracks value and VAT that is almost indistinguishable from standard sales tax with almost every tax code revision; to the point where it is now basically impossible to discern the actual problem the perennial tax code reforms are meant to resolve.

A missed opportunity

It is already evident, even before deductions start in May 2022, that e-levy administration, because of fundamentally sub-optimal design, will be faced with similar confusions.

For instance, no one has factored the complexity of managing exemptions for things like loan repayments in a country where hire-purchase, discounting, supplier credit of various shapes and forms, and payday loans are surging in popularity but accounting systems across the board remain rudimentary.

The burden will now be on fintech operators to implement complicated regimes across networks and service platforms to detect when a payment is for payroll, loan servicing or cross-network same-owner wallet top-up, and exempt it from the charge – a massive engineering challenge when one remembers that senders and recipients will rarely be on the same network. One can only guess the impact on user experience and customer service, and the impetus towards cash as a substitute.

Given that many fintech users currently make digital payments for offline rather than online goods and services, recruitment of the next generation of digital economy consumers could suffer considerable setbacks due to the levy’s contribution to price margins and to impediments created by increased KYC (know-your-customer) rules and requirements introduced to enable tax compliance by digital service operators/providers.

Meanwhile, the arbitrariness of the exemption regime means that in coming months and years, any group with sufficient political capital shall be trooping to politicians to demand their own special treatment.

All in all, the knee-jerk approach to introducing a tax with such potentially systemic effects has robbed the country of the opportunity to carefully design things in a way that would have balanced the legitimate fiscal needs of the state, with the need to maintain positive consumer sentiment.

The government will soon find that revenues extracted from e-levy may well be offset by losses elsewhere in the tax portfolio. If this were to happen, the fiscal consolidation it is so eager to sell to international investors, so that it can return to its expensive borrowing habits in the international capital markets, may in such circumstances prove somewhat elusive.