The Finance Act amendments quietly published in the Gazette this March have made three material changes that affect every Nigerian SME — and almost no local accountant we have spoken with this quarter has fully internalised them yet.

This is not a comprehensive technical analysis. It is the operational note we sent to every Nigerian client in our practice last week. If you are running a business in Nigeria, these three changes affect you whether you have noticed them or not — and the penalty math, as ever with FIRS, is unforgiving.

Change one: the withholding tax compliance window has tightened

The most consequential change. Previously, WHT remittance was due by the 21st of the month following the deduction. Under the amendment, this is now the 10th of the month following — a full eleven days earlier — and the penalty regime for non-compliance has been restructured.

The new penalty structure compounds faster than the old one. Where the previous regime applied a flat 10% penalty plus interest on the principal amount, the amended regime applies:

• 10% on day one of default
• An additional 1% per day of continuing default, capped at 30%
• Interest at CBN MPR plus 5%, compounded monthly rather than simple

For a business that misses a single WHT cycle on a NGN 5 million payment, this is the difference between a NGN 500,000 exposure under the old rules and a NGN 2.1 million exposure under the new ones — assuming an average four-month resolution timeline, which is roughly what we see in practice.

"The amendment moves the WHT deadline from day 21 to day 10. Eleven days. Most CFOs do not know yet. And the penalty math is brutal."

Change two: VAT input claim deadlines now expire faster

The window during which a registered business can claim input VAT against output VAT has been shortened from 18 months to 12 months from the date of the underlying invoice. This is buried in the section on input VAT recovery and is easy to miss.

For most SMEs, this is operationally invisible — invoices are processed within weeks, and input VAT is claimed in the same return cycle as the output VAT it offsets. But for businesses with longer payment terms, project-based billing, or invoices held in dispute, this creates a hard expiry on previously deferred input VAT.

We have already identified, in three client diagnostics this quarter, businesses sitting on aged input VAT balances that will be permanently lost if not claimed within the next two return cycles. The pattern repeats: invoice issued in 2024, payment disputed, claim deferred, dispute resolved in 2025, and now — under the amended rules — the input VAT claim window may already have closed.

What to do this quarter

If you have any aged trade payables or any input VAT pending dispute resolution, the action is simple but time-sensitive:

1. Run an aged input VAT report from your accounting system today
2. Identify every invoice older than nine months where input VAT has not yet been claimed
3. Reconcile the underlying transactions and supporting documentation before the 12-month window closes
4. File the input VAT claim in the next return cycle, even if the cash payment is still outstanding

Change three: digital services now reportable separately

The third change affects a smaller cohort but more deeply. Businesses receiving digital services from non-resident providers — software subscriptions, advertising platforms, cloud infrastructure, payment processing, even some courier APIs — must now report these separately under a new digital services schedule appended to the VAT return.

This is the operational manifestation of Nigeria's earlier shift to taxing digital services from non-resident providers. Where previously the obligation sat with the non-resident to register and remit, the new schedule requires the Nigerian recipient to confirm the provider's registration status, the input VAT claimed, and any reverse-charge VAT calculated.

For most SMEs we work with, the practical effect is that a list of recurring foreign-currency software subscriptions — AWS, Google Workspace, Stripe, Slack, Notion, Cloudflare — now needs to be tracked at the line-item level on every VAT return.

Who is most affected

• Tech-enabled SMEs — companies with heavy cloud and SaaS spend
• E-commerce businesses — using foreign payment processors and advertising platforms
• Agencies and consultancies — with international software stacks
• Anyone using foreign-platform marketing — Meta, Google, LinkedIn ad spend

The reporting itself is not onerous once the data is structured. The problem is that most chart-of-accounts setups we encounter do not currently segment digital services from other foreign-currency expenses. Restructuring this is a one-time effort; not doing it means manually reconstructing the data every month.

What this means in aggregate

Each of these three changes is, individually, manageable. The compounded effect is that the Nigerian compliance environment in 2026 is materially less forgiving than it was in 2025 — and the gap between what compliant operators do and what most SMEs are actually doing has widened.

We are running compliance refresh diagnostics for every Nigerian client in the practice this quarter. The pattern, even among well-run businesses, is consistent: between two and four genuine exposure points per business, almost always identifiable inside an hour of review.

If you are running a Nigerian SME and your accountant has not raised these three changes with you this quarter, that is worth a conversation. Not because we want to sell you an engagement, but because the cost of getting these wrong has gone up — and the cost of getting them right has not.

One last thing

The amendments are buried in a Finance Act that runs to 84 pages. We have produced an internal one-page summary covering all material changes affecting SMEs — the three above and a further seven of more limited application. If you would like a copy, it is available on request to clients and non-clients alike. Email us at info@athenataxlegaladvisory.org with the subject line "FIRS 2026 summary" and we will send it within the working day.

If you run a business in Kenya, Uganda, or Tanzania and your accountant tells you M-Pesa is reconciled, there is a very good chance it is not. This is not because your accountant is dishonest. It is because the standard bookkeeping playbook does not work for mobile money — and almost nobody in the local profession has rewritten the playbook.

We have onboarded forty-plus clients onto our M-Pesa reconciliation methodology in the past two years. In thirty-eight of them, the previous accountant's reconciliation was materially incomplete. The exposure ranged from KES 30,000 of misclassified transaction fees to a six-figure cash position that did not exist on the balance sheet. The pattern is consistent enough that we now treat M-Pesa reconciliation as a discipline of its own.

This memo is the methodology we use, written up for any operator or in-house accountant who wants to fix this without engaging us. If you read it and decide to do it yourself, that is a good outcome. If you read it and decide it is too much work to do every month, that is also a good outcome — and you know what to outsource.

The bank-statement reconciliation that does not work

The standard bookkeeping approach treats M-Pesa as if it were a bank account. The accountant downloads the M-Pesa statement, matches the credits to recorded sales and the debits to recorded payments, and signs off when the closing balance agrees to the M-Pesa balance.

This works when M-Pesa is used for occasional, low-volume transactions. It fails almost completely for businesses where M-Pesa is a primary collection channel — which is most retail, hospitality, transport, and small-format service businesses in our markets.

The failure modes are predictable. M-Pesa transaction fees get bundled into the deposit amounts and never separately recognised. C2B paybill numbers receive payments tagged to customer references that the bookkeeper cannot match to invoices. Reversals and refunds happen days after the original transaction. Float top-ups from the operating account get treated as expenses instead of asset transfers. The cumulative effect is books that look reconciled but are quietly wrong by a few percent every month — which is exactly the range that does not get noticed until it has compounded into something significant.

"In 38 of 40 client diagnostics this year, the previous M-Pesa reconciliation was materially incomplete. The pattern is consistent. The fix is not glamorous, but it works."

The four layers of a proper M-Pesa reconciliation

What follows is the methodology we run monthly for every client with material M-Pesa activity. It is layered because each layer catches errors that the layer above misses.

Layer one: the statement-to-ledger reconciliation

This is the layer most accountants stop at. Match the M-Pesa statement transactions to the ledger entries. If the totals agree and individual entries cross-reference, the reconciliation is "complete".

It catches transposition errors, missed entries, and gross misclassifications. It does not catch fee leakage, reversal mismatches, or counterparty discrepancies. Useful as a foundation, insufficient on its own.

Layer two: the gross-to-net reconciliation

For every M-Pesa transaction, there is a gross amount (what the customer paid or what you sent) and a net amount (what hit your float account after fees). These differ by the transaction fee — usually a small amount, but cumulatively material.

The reconciliation here is straightforward but rarely done properly:

• Gross sales recognised on the income statement should equal sum of gross M-Pesa receipts
• Transaction fees should be recognised as a separate expense line, not netted off revenue
• Float account balance should equal net receipts minus net payments minus closing fees

When this layer is skipped — which is most of the time — revenue is understated, the fee expense is hidden, and the operating margin looks better than it actually is. For a retail business with KES 8 million of monthly M-Pesa turnover and standard B2C transaction fees, the misrecognition is in the order of KES 30,000 to KES 50,000 every month. Annualised, it is material.

Layer three: the counterparty reconciliation

This is where it gets harder, and where most reconciliations break down entirely. The M-Pesa statement shows you that you received KES 4,500 with reference "JS123". Your sales ledger shows an invoice to John Smith for KES 4,500 with reference "JS123". Easy match.

Except your sales ledger also shows another John Smith invoice for KES 4,500 from the previous week with reference "JOHN-001". And M-Pesa has two more deposits this month with reference "JS123" that should be applied to two different invoices. And then there is a deposit with no reference at all — just "KES 4,500 from 0712XXXXXX".

Counterparty reconciliation requires matching M-Pesa receipts to specific customer ledger entries — not just by amount but by reference, by date, by phone number, by context. For paybill businesses with hundreds of monthly transactions, this is where automation pays for itself. For till-based businesses, manual review of unmatched items is the only option.

The error we see most often at this layer: customer accounts that show as overpaid (because two payments were applied to one invoice) sitting alongside accounts that show as overdue (because the second invoice never got the matching payment). The net cash is reconciled but the receivables ledger is wrong, and the wrong customer gets the dunning letter.

Layer four: the timing reconciliation

M-Pesa transactions are real-time. Bank settlements from Safaricom to your business bank account are not — there is typically a one-day delay, sometimes longer over weekends and holidays. Reversals and disputes can settle days or weeks after the original transaction.

The timing reconciliation tracks:

• In-transit balances — M-Pesa receipts that have been credited to your till but not yet swept to your bank account
• Pending settlements — funds Safaricom is holding pending settlement, particularly around month-end
• Reversal positions — transactions reversed in the current month that relate to prior-month sales
• Float top-ups — transfers from the operating account to the M-Pesa float, which should be recognised as asset transfers not expenses

This is the layer that ensures the month-end M-Pesa balance on the balance sheet reflects what is actually recoverable, not what the statement happens to show on the last day of the month.

The compounding effect

None of this is glamorous. None of it makes the financial statements look more impressive. What it does is make them correct — which matters when you are pricing the business, defending an audit, raising from a DFI, or simply trying to know how much cash you actually have.

What we recommend

If you are running an in-house finance function with material M-Pesa activity, three concrete steps:

1. Pick a recent month and run all four layers manually. You will find errors. You will also find that the work is not as bad as it sounds, once the methodology is internalised.
2. Build it into the monthly close cycle. The longer you leave gaps between reconciliation cycles, the harder it gets to unwind discrepancies.
3. Consider whether the volume justifies automation. For paybill businesses above roughly KES 5 million monthly M-Pesa turnover, a proper reconciliation tool pays for itself.

If you are not running this kind of reconciliation today, the most likely scenario is that your books are wrong by a small percentage every month and it has not yet caused a visible problem. The problem becomes visible when you try to raise capital, sell the business, or get audited — at which point it is too late to do anything about it cheaply.

Every fundraise we have supported across the continent in the past three years has slowed down at the same point — the data room. The pitch deck lands well. The fund expresses interest. The term sheet circulates. And then diligence opens and the requests start piling up, and the founder spends six weeks pulling together documents that should have been pulled together in advance.

The cost is real. Diligence delays slow term-sheet conversion. Slow conversion gives competitor processes time to start. Founders who are six weeks behind their own diligence schedule get worse terms, more often. We have watched this happen to good companies repeatedly. It is avoidable.

This memo is the checklist we use when we are engaged as the financial workstream for a fundraise. Build this set of documents now — before you have a term sheet, before you have a fund expressing interest, before you even know which round you are running — and diligence becomes a week, not six. Read it through, decide what is missing in your current set-up, and start on the gaps this weekend.

The seven sections of a complete data room

A serious data room has seven sections. Funds will ask for everything in all seven, but they will ask for them in roughly this order. Build them in this order.

Section one: corporate and legal

This is the section funds work through first because it gates everything else. If your cap table is messy or your incorporation documents are incomplete, no amount of strong financial performance will move the deal forward.

• Certificate of incorporation and all share-allotment certificates
• Memorandum and articles of association, including all amendments
• Current cap table, with any options, warrants, or convertible instruments included
• All shareholders' agreements and related contractual arrangements
• Board minutes for the past 24 months
• Statutory registers — register of members, directors, charges
• Any subsidiaries or related-party entities, with their respective constitutional documents

The most common gap we see: a cap table that includes informal commitments — early-stage employees promised options, advisors promised equity, friends who wrote cheques without documentation. If any of these exist in your business, get them documented in writing before diligence opens. Funds will not invest where the equity story is unclear.

Section two: financial statements and accounting policies

The core financial section. This is where IFRS readiness matters — funds will expect at minimum IFRS for SMEs presentation, and most DFIs will require full IFRS. If your books have been kept on a cash basis, this is the place where the conversion work hits.

• Audited or reviewed financial statements for the past three years, where available
• Management accounts for every month of the current year
• IFRS or IFRS for SMEs accounting policies documented in writing
• A clean trial balance reconciling to the management accounts
• Aged debtors and creditors listings
• Bank, mobile money, and payment processor reconciliations
• Fixed asset register with depreciation schedules

"The single highest-impact piece of work an African SME can do before fundraising is to convert from cash to accrual basis. It is unglamorous. It is also non-negotiable for serious capital."

Section three: tax and statutory compliance

Funds care about tax exposure because unidentified tax liabilities can change the economics of an investment overnight. They will ask for everything that demonstrates compliance.

• All tax returns filed in the past three years — corporation tax, VAT, WHT, PAYE
• Tax Compliance Certificate or Tax Clearance Certificate, current and valid
• Correspondence with the tax authority on any open matters
• Any tax assessments, demand letters, or audit findings, with status
• Statutory deduction remittances — NSSF, NHIF, pension, NSITF, UIF — with proof of payment
• Transfer pricing documentation if cross-border
• Any tax incentives or holiday certificates you are relying on

The single hardest thing to retrofit in this section is a clean TCC. If yours has lapsed or is held up, get it current now. Funds will ask. There is no satisfactory answer for "it is in process".

Section four: revenue, customers, and commercial

This is where funds form their view on the business itself — not just the financial reporting, but the underlying commercial reality.

• Top 20 customers by revenue, with concentration analysis
• Customer contracts for top 10, with attention to termination clauses
• Revenue recognition policies for each major product line
• Cohort analysis showing customer retention and revenue expansion
• Gross margin by product line and customer segment
• Sales pipeline with weighted probabilities
• Any churn analysis, NPS data, or customer satisfaction metrics

For African SMEs the customer concentration question is particularly delicate. Funds expect concentration in earlier-stage businesses. They want to see that you understand the concentration risk and have a plan to diversify — not that you have already diversified.

Section five: operations and personnel

The operational due diligence is often the section where less-prepared businesses lose time. Funds will want to understand how the business actually runs.

• Full org chart with reporting lines
• Employment contracts for senior team, including all option arrangements
• Payroll register for the past 12 months
• Material supplier contracts and concentration analysis
• Property leases for all offices and operating premises
• Any IP registrations — trademarks, patents, software ownership
• Business continuity, disaster recovery, and information security policies

Section six: forecasts and the business model

The forward-looking section. Funds invest in the future, not the past — and they want to see how you think about that future.

• Three-year operating model with monthly granularity for year one
• Unit economics by product, channel, or customer segment
• Capital plan — what the use of proceeds will be, with milestones
• Sensitivity analysis on key assumptions
• Cash runway under base case, downside, and stress scenarios
• Headcount plan tied to revenue assumptions

The model needs to be a model, not a spreadsheet. Drivers should be visible and changeable. Assumptions should be documented separately. The output should reconcile to the financial statements in the past and to the cash runway going forward. Funds will run their own sensitivities — make sure your model can take them.

Section seven: governance and risk

Often the section least developed in African SMEs we work with. Funds will want to see that you have started to think about risk management as the business scales.

• Risk register with mitigation status for material risks
• Board composition and any committee structures
• Internal controls documentation — particularly cash controls
• Insurance coverage — public liability, professional indemnity, key person
• Material litigation, threatened or actual, with status
• Anti-bribery, anti-corruption, and AML policies, where applicable
• ESG framework, where the fund is impact-focused

What we do for our Fractional CFO clients

For every business in our Fractional CFO tier where fundraising is an active consideration, we maintain the data room continuously — not as a project that gets initiated when a term sheet appears, but as a discipline that runs in parallel with monthly operations. New customer contracts get added. New employment contracts get added. Audited statements replace reviewed ones when they become available. Quarterly board minutes get filed.

The result is that when diligence opens, the work is to grant access and respond to questions — not to spend six weeks assembling the materials. The fundraise closes on schedule. The founder retains negotiating power. Better terms get achieved.

If you are 12 to 18 months away from a planned raise and your data room does not look like the list above, the time to start is now. The work is not glamorous. The compounding effect is real.

Of every ten tax diagnostics we run, eight will surface a material withholding tax issue. The numbers vary — sometimes it is a few months of missed remittances, sometimes it is several years of compounded exposure — but the pattern is so consistent we now treat WHT as the first area to review whenever we onboard a new client.

The reason it goes wrong is not complicated. WHT sits at the intersection of three things that small businesses are bad at handling well: tax-authority deadlines, supplier-payment processes, and the discipline of separating gross from net amounts. When it fails, it fails quietly — for months, sometimes years — until something forces a reckoning.

Why WHT is the deduction nobody catches

Three structural reasons.

First, WHT is borne by the payee, not the payer. When you fail to remit PAYE for your employees, the employees notice — their payslips do not match their statutory deductions. When you fail to remit VAT, the tax authority notices — they expect monthly returns. When you fail to remit WHT, the supplier you paid does not notice (they got their cash) and the tax authority does not notice (until they audit). The feedback loop is broken.

Second, WHT lives in the AP process, not the payroll process. Most SMEs assign tax compliance to the payroll cycle — PAYE, NSSF, NHIF run monthly with payroll. WHT lives somewhere else: every time you pay a service provider, a landlord, a consultant, or a contractor, you should be deducting and remitting. Almost no SME has a payment process that enforces this discipline.

Third, WHT rates and thresholds change frequently. The rates in Kenya, Nigeria, and Ghana have all been amended in the past three years. Local accountants who learned WHT a decade ago apply rates that no longer exist. By the time the next FIRS or KRA audit happens, the cumulative under-remittance can be substantial.

"WHT failures compound silently. You do not notice them this month. You do not notice them next month. You notice them three years later when the demand letter arrives — and the assessment can be larger than retained earnings."

The payments that should be triggering WHT

The exact rates and thresholds vary by jurisdiction, but the general pattern across Kenya, Nigeria, and Ghana is consistent. WHT applies to:

• Professional fees — lawyers, accountants, consultants, doctors, architects
• Management and technical fees — to local and foreign service providers
• Rent — commercial and residential, above the de minimis threshold
• Contractor payments — construction, IT services, maintenance
• Director fees — separate from payroll, where directors are not employees
• Commissions — to agents, intermediaries, brokers
• Royalties and licence fees — including software licences in some jurisdictions
• Interest payments — particularly to related parties or foreign lenders
• Dividends — though here the rate and treatment varies more

For most SMEs, the largest exposures come from professional fees, rent, and contractor payments. These three categories are where the silent compounding happens.

The arithmetic of compounding exposure

Take a Nigerian SME paying KES 500,000 per month in professional fees and contractor payments combined. Apply the prevailing WHT rates — typically between 5% and 10% depending on the nature of the payment and the residence status of the payee. The monthly under-remittance, if WHT is not being deducted, is somewhere between NGN 25,000 and NGN 50,000.

Over a year, that is between NGN 300,000 and NGN 600,000 of unremitted WHT. Over three years — which is roughly the window FIRS will look at in a routine audit — it is between NGN 900,000 and NGN 1.8 million of principal.

Then add penalties. Then add interest, compounded monthly. A NGN 1.8 million principal exposure can resolve at NGN 4 million or NGN 5 million by the time the assessment is concluded.

And that is just one category of payments, at modest volume, for a relatively small business. Scale the numbers up and the WHT shortfall can exceed the entire retained earnings of the business — which is why we have seen businesses end over WHT issues that nobody noticed for years.

The four-step fix

Whether you are running an in-house finance function or working with a local bookkeeper, fixing the WHT problem is mechanical. Four steps, in order.

Step one: map your payment categories

List every category of payment your business makes — every line that hits your bank account or M-Pesa float as an outflow. Categorise each as either subject to WHT or not subject to WHT. For each WHT-subject category, identify:

• The applicable rate
• The de minimis threshold below which WHT does not apply
• Whether the payee is resident or non-resident
• The remittance deadline

This is a one-time exercise that takes a few hours. Most SMEs we onboard have never done it.

Step two: build deduction into the AP process

The only way to make WHT work consistently is to make deduction part of the payment workflow itself. Every supplier invoice should be reviewed for WHT applicability before payment is processed. Where WHT applies, two amounts are paid: the net to the supplier, and the gross-up to the tax authority.

This is operationally inconvenient. It is also the only way the discipline holds up over time. Memory does not work. Quarterly checks do not work. The deduction has to happen at the moment of payment.

Step three: monthly remittance, on the deadline

Aggregate all WHT deductions made during the month and remit on or before the deadline. The deadlines vary by jurisdiction and were tightened in some markets in 2026 (see our memo on the FIRS 2026 amendments). Late remittance attracts the same penalty regime as PAYE — punitive in all three jurisdictions.

Step four: annual reconciliation

At year-end, reconcile total WHT deducted to total WHT remitted, by payee. Issue WHT credit notes to every payee — they will need them to claim against their own tax liabilities. File the annual WHT reconciliation return with the tax authority.

What to do if you discover historical exposure

If you read this and recognise that you have a historical WHT issue — months or years of payments where the deduction was not made — the right path forward depends on the size of the exposure and the time elapsed.

For smaller exposures, voluntary disclosure to the tax authority typically results in penalties being waived or substantially reduced. Tax authorities across the region have voluntary compliance programs that reward proactive disclosure. The exposure becomes the principal plus modest interest, rather than the principal plus full penalties.

For larger exposures or where the elapsed time is significant, voluntary disclosure should be made through professional representation. The framing of the disclosure, the supporting documentation, and the timing of remittance all affect the eventual settlement. This is one of the engagements where having an experienced tax practice — ideally one with former tax-authority experience on the team — pays for itself many times over.

What you should not do is ignore it. WHT exposure does not age out, and the longer it sits unaddressed, the more it compounds. The audit will come eventually. The settlement is always better when initiated proactively than when responded to.

The African Continental Free Trade Agreement is one of the most significant trade developments on the continent in our lifetimes. It is also, from a tax structuring perspective, the source of the most consistent advisory work we have seen in the past 18 months. Operators who expanded under AfCFTA without thinking about the tax consequences are now learning, often expensively, that the trade agreement does not exempt them from the tax codes of the markets they have entered.

The single most important concept in cross-border tax — and the one most consistently misunderstood by founders we work with — is permanent establishment, or PE. This memo is what we tell our Fractional CFO clients about PE risk under AfCFTA. It is not legal advice; it is operational guidance based on what we have seen go wrong.

What permanent establishment actually means

The principle is straightforward. If your business has a "permanent establishment" in another country — broadly, a fixed place of business or persistent commercial activity there — then that country has the right to tax the profits attributable to that establishment. This applies even if you have no legal entity in that country.

The complication is the definition. Under most African tax treaties and domestic codes, PE can arise from any of the following:

• A fixed place of business — office, branch, factory, warehouse
• A construction site or project that lasts more than a threshold period
• Habitual exercise of an agent who has authority to conclude contracts on your behalf
• Provision of services in the country for more than a threshold number of days
• In some jurisdictions, significant economic presence even without physical presence

The thresholds vary. Some treaties set service-PE at 183 days in any 12-month period. Some domestic codes are more aggressive. The point is that PE is easier to trip into than most founders realise — and once it is triggered, the consequences extend backward, not just forward.

The four PE patterns we see most often

Pattern one: the sales agent who became a permanent establishment

The most common pattern. A Kenyan business expanding into Uganda hires a local sales agent — often on a commission basis, often informally — to sell into the Ugandan market. The agent meets customers, negotiates terms, "in principle" closes deals, and forwards the paperwork to the Kenyan parent for formal signature.

Under most treaty definitions and under Ugandan domestic law, this is exactly the kind of arrangement that creates an agency PE. The fact that the agent does not have formal contractual authority does not matter — what matters is whether they habitually exercise authority in substance.

The fix: either formalise the agent's role in a way that clearly limits their authority (which usually means losing their effectiveness), or accept that PE has been created and structure the business around it. The middle ground — pretending the PE does not exist — is what creates the diligence finding that kills the next fundraise.

Pattern two: the service contract that ran too long

The second most common pattern. A Nigerian SaaS business signs a multi-year implementation engagement with a Ghanaian customer. The implementation requires Nigerian engineers to spend time in Accra — sometimes a week at a time, sometimes a month — with cumulative presence over the contract life exceeding the service-PE threshold.

Ghana's threshold, like most in the region, is 183 days in any 12-month period for services-PE. If you have engineers, project managers, or consultants in country across the year, those days add up. They do not need to be consecutive. The PE triggers.

Pattern three: the warehouse that was always meant to be temporary

Specific to physical-product businesses, particularly under AfCFTA's tariff regime. A South African manufacturer expanding into Kenya leases warehouse space in Nairobi for "temporary" inventory positioning to serve East African customers. The lease starts as six months. It gets extended. The warehouse becomes a permanent feature of the supply chain.

Whether this constitutes a fixed-place-of-business PE depends on the nature of activities at the warehouse. Pure storage, in most treaty contexts, falls under the preparatory and auxiliary exception and does not constitute PE. Storage combined with order fulfilment, customer service, or any sales activity does constitute PE.

The fine line: where is the boundary between "storage" and "fulfilment"? The answer depends on the jurisdiction, the treaty, the specific facts. Operators expanding goods-trade under AfCFTA need to know where this line sits for their specific structure.

Pattern four: the significant economic presence rule

Newer, less talked about, increasingly important. Some African jurisdictions have introduced "significant economic presence" or SEP rules that can create taxable presence even without traditional PE attributes. Kenya, Nigeria, and Tanzania have all introduced versions of these rules in recent years.

SEP typically applies to digital businesses serving local customers — even where the business has no physical presence, no agent, no employees in country. Revenue from local customers above a threshold creates taxable nexus.

For African SaaS, fintech, e-commerce, and digital media businesses expanding cross-border, SEP is the structural risk that did not exist five years ago and now does. It cannot be avoided through traditional PE planning because it is not a traditional PE concept.

"AfCFTA opened the door. The tax authorities walked through it. The asymmetry — easier trade, unchanged tax rules — is where most of our cross-border advisory work originates."

How to structure around PE risk

There is no universal answer to PE risk because the right structure depends on the business model, the markets involved, the volume of cross-border activity, and the operating model. But there are four structural approaches that recur in our advisory work.

Approach one: the local subsidiary

The simplest and most defensible approach. Establish a wholly-owned subsidiary in each market where you have material activity. The subsidiary becomes the taxpayer in that market. Cross-border activity becomes intercompany activity, governed by transfer pricing rules.

This is the right approach when in-country activity is going to be material and recurring. It carries the highest setup and ongoing administration cost, but it is also the structure that holds up best under audit.

Approach two: the regional holding structure

For businesses operating across multiple African markets, a holding company structure — typically in Mauritius, Rwanda, or a comparable jurisdiction with favourable treaty networks — can rationalise the tax position across the group.

This is more involved and more expensive to set up. It is also the structure most likely to attract attention from tax authorities, which is why it needs to be done with substance — real management, real decisions made at the holdco level, real reason to exist.

Approach three: the secondment model

For service businesses where employees need to spend time in foreign markets but the underlying engagement is short-duration, formal secondment arrangements with local employers (often customers) can avoid creating PE for the home-country entity.

This works in specific circumstances and requires careful contracting. Done well, it solves the service-PE problem cleanly. Done badly, it creates employment law issues in addition to the tax issues it was meant to solve.

Approach four: the genuine restructure

For businesses that have already created PE exposure they did not intend, the right answer is often to restructure proactively. Convert the de facto presence into a formal one. Establish the entity. Pay the tax going forward. Negotiate a voluntary settlement of historical exposure where possible.

This is harder and more expensive than getting the structure right from the start. It is also frequently necessary, because most growth-stage SMEs we work with discover PE exposure only after it has been accumulating for some time.

What to do this month

If you are running a business with any cross-border activity in Africa, three concrete actions:

1. Map your cross-border activity. Which markets do you have customers in? Which markets do your employees travel to? Which markets do you have agents or contractors operating in? Most founders cannot answer all three questions accurately without checking.
2. For each market, identify the PE thresholds — both treaty and domestic. Where you are close to the threshold or above it, treat that as a priority for structuring review.
3. Pay particular attention to SEP rules in Kenya, Nigeria, and Tanzania. Digital revenue is the area where the rules have changed most quickly and where exposure is least visible.

AfCFTA is a genuine opportunity. It opens markets. It reduces tariffs. It simplifies a lot of the practical complexity of intra-African trade. What it does not do — and is sometimes confused as doing — is harmonise tax rules. The tax codes of the 54 member states remain distinct, and operators who expand under AfCFTA without engaging with this reality are the source of most of our cross-border restructuring work.

If you are running an African SME and you are reading this, there is a good chance your books are on a cash basis. Maybe formally — your QuickBooks file is set to record-on-payment, your bookkeeper recognises revenue when customer money lands. Maybe informally — the books are nominally accrual but in practice nothing gets recognised until it is paid, and accruals at year-end are reconstructed for tax purposes rather than maintained continuously.

This is fine, until it is not. The moment it stops being fine is usually the same moment: a DFI, fund, or commercial lender opens diligence and asks for IFRS or IFRS for SMEs financial statements. At that point, the conversion becomes a project — sometimes a long project, often more expensive than founders expect.

This memo is the conversion playbook we have run for 40-plus clients in the past three years. If you are anywhere within 18 months of a planned raise, read it now and start the conversion before the term sheet appears.

What the conversion actually involves

Converting from cash basis to IFRS for SMEs is not, despite how it sometimes feels, a software change. It is a methodology change. Six main areas need to be addressed.

Revenue recognition

Under cash basis, revenue equals cash received. Under IFRS, revenue is recognised when control of goods or services transfers to the customer — which may be before, at, or after the cash receipt depending on the nature of the contract.

For most SMEs, the practical effect is that:

• Sales invoiced but not yet paid become accounts receivable, with revenue recognised when invoiced (or when delivered, depending on terms)
• Customer deposits and prepayments become contract liabilities, with revenue deferred until performance
• Subscription and recurring revenue gets spread evenly across the subscription period, not lumped at receipt
• Multi-element arrangements need to be unbundled and recognised per performance obligation

This is the area most SMEs find hardest. The cash-basis instinct — "we got paid, so we earned it" — has to give way to a more disciplined view of when economic value transfers.

Expense recognition

The mirror of revenue recognition. Under cash basis, expenses equal cash paid. Under IFRS, expenses are recognised when incurred — matched to the period they relate to.

• Supplier invoices received but not yet paid become accounts payable
• Rent paid quarterly in advance gets spread evenly across the quarter
• Insurance premiums get amortised across the policy period
• Year-end accruals — bonuses, leave, audit fees — need to be properly recognised

Fixed assets and depreciation

Under cash basis, large asset purchases often hit the P&L in the period of payment. Under IFRS, they are capitalised on the balance sheet and depreciated over their useful lives. This requires:

• A capitalisation policy — what threshold differentiates an expense from an asset
• Useful-life policies by asset category
• A fixed asset register with cost, accumulated depreciation, and net book value tracked
• Disposal accounting when assets are sold or written off

Inventory

For businesses with physical inventory, cash-basis books often treat purchases as direct expenses. Under IFRS, inventory is held on the balance sheet at lower of cost and net realisable value, and only recognised as cost-of-sales when sold. This requires periodic stock counts, valuation methodology (FIFO is most common for SMEs), and a discipline around year-end stock takes.

Foreign currency

The area where African SMEs most consistently get caught. IFRS requires foreign-currency balances to be revalued at each reporting date — typically monthly — with gains and losses recognised in profit or loss. The cash-basis approach of recording everything at transaction rate and never revisiting it does not survive IFRS conversion.

Tax

Under cash basis, the tax line is whatever you remitted. Under IFRS, current and deferred tax need to be properly recognised — including temporary differences between accounting and tax treatment of revenue, depreciation, provisions, and accruals.

"The single highest-impact piece of work an African SME can do before fundraising is to convert from cash to accrual basis. It is unglamorous. It is also non-negotiable for serious capital."

The conversion timeline

For a relatively simple SME — single entity, single jurisdiction, modest transaction volume — a clean conversion takes four to six weeks. For more complex businesses — multi-entity, multi-currency, inventory-heavy — it takes two to four months.

The phases, in our experience:

Phase one: policy and chart of accounts (week 1-2)

Document the accounting policies you will apply. Restructure the chart of accounts to support accrual reporting — proper segregation of revenue categories, payables, accruals, deferred revenue, fixed assets. Without this foundation, the rest of the conversion does not hold.

Phase two: opening balance sheet (week 2-4)

Construct the opening IFRS balance sheet as at the conversion date. This involves identifying and quantifying every item that exists under accrual basis but did not appear under cash basis: receivables, payables, prepayments, accruals, the proper carrying value of fixed assets, inventory at IFRS valuation. This is the hardest phase. Get it wrong and every subsequent period inherits the error.

Phase three: comparatives (week 4-8)

Restate prior-period financials to IFRS basis. For DFI fundraising, comparatives of the previous two financial years are typically required. The work involves replaying the year's transactions on accrual basis — recognising revenue when invoiced rather than when paid, accruing expenses to the right period, depreciating fixed assets, revaluing foreign-currency balances.

Phase four: current period and ongoing (week 6 onwards)

Move the live books onto the new basis. New transactions are recorded under IFRS principles from day one. Monthly close incorporates the additional discipline — accrual reviews, prepayment amortisation, FX revaluation, depreciation runs.

What it costs

For a typical SME with two to three years of historical books to convert, the all-in cost of a clean conversion — whether done in-house or through a firm like Athena — is typically equivalent to four to eight months of ongoing accounting fees. So if you pay USD 1,500 per month for your accounting today, expect to invest USD 6,000 to 12,000 in the conversion.

The cost varies with:

• Volume of historical transactions — more transactions, more work
• Complexity of revenue model — subscription and multi-element arrangements take longer
• Inventory complexity — physical inventory adds significant work
• Number of entities and currencies — each adds complexity
• Quality of historical records — clean records convert faster than messy ones

The cost is one-time. The ongoing accrual-basis bookkeeping is typically 15 to 30% more expensive monthly than cash-basis, reflecting the additional discipline required.

The pitfalls we see most often

Pitfall one: opening balance sheet errors that propagate

The opening balance sheet is the foundation. If receivables are misstated by 10%, every subsequent period's revenue and cash position is off. We have seen conversions where this happened and was not caught for two years — at which point the audit found it and the comparatives had to be restated.

Pitfall two: revenue recognition policy that does not match contracts

Founders sometimes describe their revenue model in marketing terms ("we charge monthly") that do not match the actual contractual terms ("annual contract paid upfront, refundable in first 30 days"). The policy needs to reflect the contracts, not the pitch deck.

Pitfall three: foreign currency treatment for monetary vs. non-monetary items

The IFRS rules distinguish between monetary items (cash, receivables, payables — revalued at each reporting date) and non-monetary items (inventory, fixed assets — held at historical rate). Getting this distinction wrong is one of the most common audit findings on first-year IFRS conversions.

Pitfall four: tax line that does not reconcile

Current tax under IFRS should reconcile to the tax return. Deferred tax should be properly computed and disclosed. Both are routine areas where conversion errors hide — often not detected until the first IFRS audit.

Pitfall five: trying to convert and run live operations simultaneously

This is the operational pitfall. Founders attempt to convert prior periods while the in-house accountant continues running cash-basis books for the current period — and then everything has to be reconciled at the end. This rarely ends well. The conversion should be done with the live books transitioning to IFRS basis simultaneously, even if the historical conversion runs in parallel.

What to do if you are 18 months from a raise

Start the conversion now. Two reasons.

First, IFRS comparatives are typically required for two to three years. If your raise is 18 months out and your current books are cash basis, you need the conversion in place 18 months ago — meaning you need to convert the current year as you go, plus restate the two prior years. Starting now gives you exactly the right window.

Second, the discipline of running IFRS books for 12 to 18 months before diligence opens is what separates clean diligence from messy diligence. A fund opening diligence on a business that has been on IFRS for six weeks knows they are looking at a recent conversion. A fund opening diligence on a business that has been on IFRS for two years sees a business that takes finance seriously.

Either you do this or you do not — there is no half-way option that survives diligence. If you are anywhere on the fundraising path, start the conversion before you need it.

The statutory deduction landscape across our markets shifted meaningfully at the start of 2026. None of the changes is individually dramatic. The combined effect, however, is that payroll systems running on 2025 rates are now non-compliant — and the businesses operating those systems are accumulating exposure they have not yet noticed.

This memo summarises the changes we are implementing for clients in Kenya and South Africa specifically. The Nigerian, Ghanaian, Ugandan, Rwandan, and Tanzanian changes are less consequential this year and will be covered in a separate memo next quarter. If you are running payroll in either Kenya or South Africa, read this carefully.

Kenya: the SHA transition and what it actually means

The Social Health Authority replaced NHIF as the statutory health insurance scheme during 2024, with a transition period that extended into 2025. By the start of 2026, the transition is no longer optional — all employers should be remitting under SHA, not NHIF.

The mechanics are different from NHIF in several ways that affect payroll calculation:

Contribution structure

SHA contributions are calculated as a percentage of gross salary rather than the NHIF banded structure. The rate is 2.75% of gross monthly salary, with no upper cap. This is significantly different from NHIF, which had a maximum monthly contribution and a banded structure.

The practical effect: high earners pay materially more under SHA than they did under NHIF. For an employee earning KES 200,000 gross per month, the contribution moves from KES 1,700 (NHIF maximum) to KES 5,500 (2.75% of gross). For a finance director on KES 500,000, the difference is even larger.

Employer contribution

Unlike NHIF, where the employer's role was largely administrative, SHA introduces a matching employer contribution in some scenarios. Confirm with your senior accountant exactly what applies to your industry and employee categories.

The dependant declaration

SHA registration requires each employee to declare their dependants, with the dependants enrolled under the employee's SHA cover. This is operational work that sits with HR, not with payroll, but it gates the deduction process. Employees who have not completed their dependant declarations cannot be properly enrolled, and the employer's obligation persists regardless.

The housing levy

The Affordable Housing Levy at 1.5% of gross salary, matched by the employer at 1.5%, has been in force since 2023 but has been subject to several rate adjustments and legal challenges. As of January 2026, the levy is settled at 1.5% from employee and 1.5% from employer, with no upper cap.

For an employee on KES 100,000 monthly gross, this is KES 1,500 from the employee plus KES 1,500 from the employer — KES 3,000 per month, KES 36,000 per year. For payroll systems that have not implemented the levy, or are computing it at a wrong rate, the back-exposure for 2024 and 2025 alone can be substantial.

NITA

The National Industrial Training Authority levy continues at KES 50 per employee per month. The rate has not changed, but enforcement has tightened — we are seeing more frequent compliance audits in 2026 than in any prior year.

South Africa: the UIF rate and the SDL refresh

South African statutory changes have been more administrative than substantive in 2026, but they still need attention.

UIF rate confirmation

The Unemployment Insurance Fund continues at 1% from employee and 1% from employer, capped at a maximum monthly earnings level. The cap was adjusted upward for 2026, which means high earners pay more than they did in 2025. Payroll systems need the updated ceiling applied — running 2025 ceilings means over-deduction from some employees and reconciliation issues at year-end.

SDL methodology refresh

The Skills Development Levy at 1% of payroll continues, but the SETA classification framework was updated for the 2026 year. Some industry classifications have shifted, which affects which SETA receives the levy contribution. This does not change the amount remitted but does affect how it is reported.

UIF declaration changes

The UI-19 declaration process has been updated for 2026, with new fields required for certain employee categories. Most payroll software providers have implemented the changes by now — if your provider has not, expect rejection of submissions until they do.

Provident Fund and Pension changes

The two-pot retirement system, which became effective in September 2024, continues to affect how provident and pension fund contributions are structured and reported. For employers who have not yet fully implemented the two-pot reporting requirements, this is the priority compliance item to close out in Q1 2026.

"None of these 2026 changes is dramatic individually. The combined effect is that payroll systems running on 2025 configuration are now structurally non-compliant — and the exposure is largest for the businesses with the highest-paid staff."

What to do this month

Three concrete actions for any business running payroll in Kenya or South Africa.

Action one: confirm the rates currently configured

Open your payroll system. Identify the rates currently configured for SHA (or NHIF — if NHIF is still showing, that itself is a problem), housing levy, NITA, UIF, SDL, and any pension or provident contributions. Compare to the 2026 rates summarised in this memo and confirmed with your tax authority.

If anything is wrong, that is the priority for the week. The longer wrong rates run, the larger the back-exposure becomes.

Action two: reconcile January and February 2026 deductions

If you have already processed two months of 2026 payroll, those months should be reconciled to the new rates. Where under-deductions have occurred, the right path forward is correction in the next payroll cycle — not a build-up of accumulated under-deductions.

Action three: confirm employer matching obligations

Several of the 2026 changes — particularly the housing levy and elements of SHA — involve employer matching contributions. The employer cost is real and should be reflected in your cost-to-company calculations, your budget for the year, and your management reporting.

What we do for our clients

For every Operator-tier and Fractional CFO client in Kenya and South Africa, the 2026 rate refresh was completed in December 2025 ahead of January payroll runs. Statutory remittance schedules were updated. Employee communications about the SHA transition were drafted and issued.

For Essentials-tier clients, the same updates are running through Q1 2026 on the standard service cadence.

If you are running payroll in-house and are not certain whether your system has been updated, the most efficient diagnostic is to take one current payslip for a senior employee and recompute every statutory deduction from first principles. If your recomputation matches the payslip, the system is on the right rates. If it does not, the difference is the exposure — and the longer it runs, the larger it grows.

Compliance with statutory deductions is one of those areas where the cost of getting it right is low, the cost of getting it wrong is high, and the cost compounds with time. Now is the right month to confirm where you stand.