It's the System": What Six Weeks for a TIN Is Really Costing Sierra Leone

An open analysis for the National Revenue Authority and the Ministry of Finance — on why a tax ID takes six weeks, what that delay costs a treasury already short of money, and how to fix it without a mega-contract.

Tabempa Engineering — Insights Digital Government by Alpha S. Mansaray

A sentence that explains everything

It took six weeks to obtain a Taxpayer Identification Number for a company in Freetown. Six weeks, for a number that the official process promises in two to three days, and that the country's own investment literature advertises as available, in some cases, on the same day.

That gap between the promise and the reality is the subject of this piece. But the more revealing thing was not the wait. It was the answer. Every time a question was put to a member of staff — where the application is, what is it waiting on, can anything move it — the reply was the same three words: "It's the system."

This article is written in good faith and addressed first to the Government of Sierra Leone, because the people who said those three words were not the problem, and treating them as the problem would be both unfair and useless. A clerk who says "it's the system" tells you something precise and important: that they have been handed a process they cannot see into, cannot override, and cannot be held responsible for. They are as trapped by it as the applicant in front of them. "It's the system" is not an excuse. It is a diagnosis — and once you read it correctly, it tells you exactly what is broken and what to do about it.

A word on method, because Tabempa's standard is evidence over assertion: the cost figures below are built from published Sierra Leonean and World Bank data, and where a number has to be estimated, the model and every assumption are shown so that any official can check the working and substitute the Authority's own inputs. We are not interested in a dramatic headline number. We are interested in a minister  who can forward to a colleague without it falling apart.

What "it's the system" actually reveals

Strip the phrase down and it exposes five distinct failures, none of which is the fault of the person at the desk.

A broken process was digitized instead of redesigned. Sierra Leone has, to its credit, invested steadily in tax technology — the Domestic Tax Information System, the VAT processing system, the Integrated Tax Administration System (ITAS) that now sits behind taxpayer registration, the electronic cash register scheme, the electronic single window for trade. This is real money and real effort. But automating a slow, multi-desk, paper-shaped process does not make it fast; it makes it a slow process with a database attached. The published route still requires a combined registration form for the Office of the Administrator and Registrar General and a separate ITAS application for the Domestic Tax Department — the same applicant data, entered into two systems, then waiting on a manual hand-off between them. That is not transformation. It is, in the well-worn phrase, paving the cow path.

There is no human override. When a case stalls, no one at the counter has the authority or the visibility to reach in and move it. The system permits exactly one path, and when reality deviates — a mismatched name, a field that won't validate, a record stuck between two departments — there is no exception desk, no empowered officer, no manual route around the blockage. So the case simply sits, and the only honest thing the staff member can say is "it's the system."

Nothing talks to anything else. A TIN application re-collects information the government already holds — from business registration, from national identity records. Because these systems are not interoperable, data is entered repeatedly, each re-entry a new chance for a mismatch that halts the file. Entering a citizen's or company's core details once and reusing them everywhere is the single highest-leverage fix in digital government, and its absence is felt in every stalled queue.

The infrastructure is assumed not engineered. Any system that depends on continuous power and connectivity is built on sand here. Sub-Saharan firms experience around nine power outages in a typical month, fewer than half of Africans have a reliable electricity supply, and in March 2024 a set of submarine-cable cuts degraded internet access across thirteen West African countries at once. A registration system with no offline capability doesn't slow down when the link drops — it stops, and the applicant is simply told to come back.

No one owns the outcome. This is the deepest failure, and the phrase names it perfectly. "The system" is a subject with no address. You cannot escalate to it, appeal to it, or hold it accountable. When the responsible party is a piece of software, responsibility has quietly left the building — and a government that cannot say who owns the time-to-issue for a TIN cannot manage it, because you cannot manage what you refuse to name.

What the delay costs the treasury

Here is the part that should command attention in the Ministry of Finance, because the irony is sharp: this is a government urgently, publicly short of revenue, running a system that throttles the very formalisation that would produce it.

The fiscal picture is not in dispute and comes from the government's own mouth. In early 2025 the Minister of Finance, relaying the President's stated disappointment with revenue performance, put the revenue-to-GDP ratio at roughly 8.1% and the Commissioner-General of the NRA acknowledged a revenue shortfall of about 25% in the opening months of that year. Sierra Leone's tax-to-GDP ratio sits far below the African average of around 16%, public debt has climbed past 90% of GDP, and more than 70% of the working population is in the informal sector — outside the tax net entirely. Every one of those numbers points the same direction: the binding constraint on the state is its inability to bring economic activity into the formal, taxable economy.

Now connect that to the queue. A business without a TIN cannot legally invoice, cannot open a corporate bank account, cannot bid for a government or NGO contract, cannot clear imports cleanly, and cannot remit the taxes the Authority is trying to collect. Every week a business spends waiting for its number is a week of formal, taxable activity that does not happen. The registration delay is not a customer-service annoyance. It is a direct break on the one thing the fiscal strategy depends on.

How large is the brake? Let us build it transparently rather than assert it.

The physical cost — which needs no money attached to be alarming.

• Assumption A: Sierra Leone registers on the order of 2,000 new formal limited-liability companies a year. (World Bank entrepreneurship data records a peak of about 1,980 in 2018; the true flow including business names and sole proprietorships is higher, so this is conservative.)
• Assumption B: the excess delay beyond the official 2–3 day target is taken here as six weeks. This is one documented case, not a measured national average — and the NRA is the only party that holds the real distribution. We invite the Authority to publish it.

Multiply them: 2,000 registrations × 6 weeks ≈ 12,000 business-weeks lost to the formal economy every year — roughly 230 business-years of formal activity sitting idle in a queue, annually. Even if the true average delay is half our assumption, that is still over a hundred business-years a year of activity the formal economy never captures. This figure uses no invented Leone amount and is hard to dispute; it is simply arithmetic on the Authority's own throughput.

The monetary cost — a formula, not a fabrication. To put money on it, the model is:

Deferred formal turnover per year = (registrations/year) × (delay in months) × (average monthly turnover of a newly-formalising business)

Deferred tax per year ≈ deferred formal turnover × effective tax take

We deliberately do not invent the average-turnover input, because the NRA holds it and a made-up figure would rightly be dismissed. Plug in the Authority's real median for newly registered SMEs and the Leone value falls out immediately. The structure is what matters: the cost scales linearly with the delay, which means halving the wait halves this loss — the fix pays for itself in recovered revenue.

The two largest costs are the ones hardest to measure. First, deterrence: an unknown but real share of would-be formal businesses, faced with a six-week black box, simply give up and stay informal — permanently outside the tax net, in a country where seventy percent of work already is. Second, trust: every applicant who endures this becomes a citizen who believes government digital systems do not work, which raises the resistance to every other e-government and tax-compliance initiative the state is trying to launch. A revenue authority's most valuable asset is voluntary compliance, and nothing erodes it faster than a system that wastes a compliant person's six weeks and then cannot tell them why.

The fix — and what it is not

The fix is not another large, monolithic, single-vendor system bought at great expense to replace the last one. That pattern is how Sierra Leone arrived here: capital spent on software wrapped around an redesigned process. What follows is incremental, affordable, and — critically — buildable and maintainable by Sierra Leonean engineers. Each item is tagged for honesty: proven (working somewhere already), plausible (works elsewhere, untested locally at scale), or aspirational (needs new capacity).

1. Redesign the process before touching the software. Map what happens to an application, step by step, and delete every step that exists only because the paper form once required it. Merge the OARG and NRA registration into a single application captured once. Proven — the same-day registration the government already advertises proves the streamlined path is possible; it simply isn't the default.

2. Build a human override and a named exception desk. Give a specific, accountable officer the authority and the visibility to resolve stuck cases, with a published escalation path and a service-level commitment. Proven, cheap, and implementable now. This alone would have dissolved the six-week wait.

3. Enter data once; reuse it everywhere. Connect registration, tax, and national identity so core details are pulled, not re-typed. A government data-exchange spine — the model pioneered by Estonia's X-Road and adopted across several states — is the durable answer. Proven elsewhere; aspirational for Sierra Leone in the near term, which is why it belongs in a phased plan, not a single leap.

4. Make issuance survive the grid. Design the registration workflow offline-first: capture and queue locally, synchronize when power and connectivity return, so an outage delays nothing. This is ordinary resilience engineering — the same queue-and-forward discipline that keeps payment systems running — and it is squarely within local capability. Plausible/proven.

5. Show the applicant where their file is. A status tracker, with updates over SMS or USSD so it reaches any phone, turns a black box into a glass one. In a country where mobile money has already put transactional messaging in every hand, this rides infrastructure that already works. Proven.

6. Publish the metric, and name the owner. Measure and publish the median time-to-issue for a TIN, and assign one accountable official to it. The moment "the system" has a human owner with a public number attached, "it's the system" stops being a place responsibility goes to disappear. Proven, and free.

7. Build incrementally and locally. Ship the cheap, high-impact items first (the exception desk, the status tracker, the single form), measure, then iterate toward interoperability. Resist the mega-contract. Systems Sierra Leoneans build and maintain are systems Sierra Leone controls. The governing principle, not a single project.

A closing note, in good faith

None of this is written to embarrass the National Revenue Authority, whose mandate — financing the country's own development from its own resources — is among the most important work in the state. It is written because the Authority's goal and the applicant's frustration point to the same fix. The NRA wants more formal businesses and paying. The six-week queue keeps businesses informal and silent. Closing that gap is not a cost to the revenue mission; it is the revenue mission.

"It's the system" was the most honest thing anyone at that counter could have said. The task now is to make it a sentence no Sierra Leonean ever has to hear again — not by blaming the people who say it, but by building a system worth defending, owned by someone who can. That is work this country's own engineers can do. Tabempa's name, in our dialect, means let's make. On this, we would be glad to help.