Tax systems are built on an uncomfortable bargain. Governments demand money from economic activity; businesses accept that obligation in return for rules that are clear, predictable and applied evenly. When that bargain breaks down, the damage is greater than the revenue that goes uncollected. It weakens competition, encourages informality and makes honest taxpayers wonder why compliance should be their burden.
Ethiopia has spent years trying to strengthen that bargain. Yet recurring disputes between tax authorities and businesses suggest that the country still has a fundamental problem: the tax administration does not always see the economy in the same way that the economy operates.
Businesses increasingly operate across jurisdictions, premises and channels. A company may be licensed in one place, keep its goods in another, sell them somewhere else and receive payment through a financial institution that has yet another institutional relationship with the state. Commercial reality is fluid. Tax administration, by contrast, can remain tied to licences, addresses and administrative boundaries.
That mismatch is becoming harder to ignore.
The recent action by the Addis Ababa Revenue Bureau against vehicle importers is only the latest manifestation. The bureau says some importers registered in other regions have been selling vehicles in the capital through commission agents and from premises that are not properly registered for such activity. The authority argues that sales conducted in Addis Ababa should generate the tax obligations applicable there.
The law must be enforced. Businesses cannot choose their tax obligations by choosing the address on a licence. If an importer is deliberately concealing sales, failing to issue receipts or avoiding taxes that are legally due, there is little ambiguity about what should happen.
But enforcement should not obscure the more important question.
Why should the tax authority have to discover a transaction through physical inspection when much of the information needed to identify that transaction already exists somewhere in the public system?
That is the question Ethiopia should be asking.
From Tax Collection To Transaction Intelligence
A modern tax administration should not depend primarily on finding businesses after they have failed to comply. It should be capable of reconstructing the economic life of a transaction.
Consider a vehicle.
It enters the country through customs. It is attached to an importer. It has a vehicle identification number. It may be moved to a warehouse or display site. It is advertised. A buyer pays for it. A receipt is issued, or should be. Ownership is eventually transferred and the vehicle is registered.
There is a trail.
The same principle applies well beyond cars. Goods enter through importers, pass through wholesalers, retailers and agents, and ultimately reach consumers. Payments pass through banks and other financial channels. Businesses obtain licences and taxpayer identification numbers. Invoices and receipts generate records.
The state does not necessarily suffer from a lack of information. It can suffer from information that remains trapped inside institutional compartments.
That is a more difficult problem than simply catching a dishonest trader.
It requires tax authorities, customs officials, licensing agencies and other relevant institutions to know enough about the same transaction to establish what happened, who was responsible and where the tax obligation arose.
Ethiopia's federal structure makes this especially important. Tax responsibilities are divided among federal, regional and city administrations. Such a structure can work, but only when information moves as easily as the economic activity being taxed.
Otherwise, administrative boundaries can become convenient blind spots.
Merkato’s Lesson Has Not Disappeared
The experience of Merkato offers an instructive precedent.
The market has repeatedly become a focal point for disputes over receipts, undocumented transactions and tax compliance. When authorities tightened enforcement, merchants raised concerns about the practical difficulties of documenting transactions and, in some cases, about the conduct of suppliers further up the chain.
The important lesson is not that merchants were right or that the tax authorities were wrong. Both propositions can be true in different cases: a trader can have a legal obligation to document a sale, while the supply chain behind that trader can also suffer from poor documentation. This distinction matters because tax compliance is not created at the point where an inspector arrives.
It is created across the transaction. A retailer cannot produce a reliable record from an upstream transaction that was never properly documented. An importer cannot credibly demonstrate complete compliance if sales records are fragmented across agents and premises. And a tax authority cannot administer a sophisticated economy effectively if it sees only the last visible link in a chain.
Merkato and the vehicle trade therefore share a more important feature than the fact that both have generated tax disputes.
In both cases, the question is whether the authorities can see the full commercial chain. Fragmented tax administration can distort competition.
A compliant business incurs the cost of registering properly, maintaining records, issuing receipts and paying the taxes due. A competitor that conceals sales or exploits administrative gaps can reduce its effective cost.
The result is perverse. The taxpayer who follows the rules becomes less competitive than the one who does not. That is why tax enforcement is ultimately an economic policy, not simply a fiscal exercise.
Yet fairness runs in both directions.
Businesses must comply with the law. They cannot invoke administrative complexity as a licence to conceal income or avoid taxes. But the state also has a responsibility to make compliance intelligible. Rules that are unclear, overlapping or difficult to administer create fertile ground for disputes and give bad actors room to hide among legitimate businesses.
A serious tax system should therefore distinguish between deliberate evasion, administrative error and genuine uncertainty. Treating all three as the same thing is poor administration.
Inspections produce immediate headlines. Sealed premises demonstrate action. Penalties create a visible deterrent. In an environment where tax collection is under pressure, these measures can produce results.
But they are an expensive way to run a tax system if they become its principal mechanism.
A government cannot inspect its way into a modern revenue system. The more durable objective is to make the transaction itself difficult to hide. That requires connecting the dots.
For vehicle imports, customs information could be reconciled with taxpayer identification, business licences, authorised premises, electronic sales records and eventual vehicle transfers. Suspicious discrepancies could then trigger targeted audits.
The same architecture could apply elsewhere. Rather than asking inspectors to search for every undeclared transaction, authorities could use data to identify businesses whose reported sales are inconsistent with imports, inventory, payments or other legally accessible information.
That would change the character of enforcement. Instead of broad inspection followed by investigation, the system could move towards risk-based investigation informed by evidence.
The difference is significant. One approach searches for violations. The other identifies where violations are most likely to occur.
Ethiopia has made digital transformation a central part of its economic reform agenda. But putting existing procedures on computers is not the same as redesigning how government understands economic activity.
A fragmented paper system can become a fragmented digital system. The objective should be interoperability. A taxpayer should have a consistent identity across relevant government systems. A business should not effectively acquire a new tax personality every time it crosses an administrative boundary. Transactions should be traceable without requiring every agency to maintain its own isolated version of the same economic reality.
This also requires safeguards. Greater use of data must come with clear rules on access, accountability, privacy and the legitimate purposes for which information can be shared. Tax intelligence should strengthen administration, not create arbitrary surveillance.
Nor should technology become a substitute for due process. If an algorithm or data mismatch identifies a taxpayer as high-risk, that should lead to scrutiny, not automatic punishment. The taxpayer must still have the opportunity to explain discrepancies and challenge decisions.
A stronger tax system is not one that gives government unlimited power. It is one that gives government better evidence.
Ethiopia's revenue needs are substantial. Public investment, social spending and debt obligations all place pressure on domestic revenue mobilisation. The temptation is therefore to focus on the size of the tax take.
But a country's tax system should be judged by more than how much money it collects. It should also be judged by how it collects it.
A system that raises revenue through unpredictable enforcement may collect money in the short term while weakening confidence in the long term. A system that makes compliance straightforward, detects evasion intelligently and applies rules consistently can gradually broaden the culture of compliance.
That is the more difficult reform. It requires less political theatre and more institutional plumbing. It means agreeing which authority is responsible for what, ensuring that information can move between them and establishing clear rules for businesses operating across jurisdictions.
It also means accepting that not every tax problem is evidence of criminal intent.
The distinction is essential for a country trying to move businesses from informality into the formal economy. If formalisation is accompanied by unpredictable costs and enforcement, businesses may have an incentive to remain outside the system. If formalisation brings clarity, access to services and a predictable relationship with the state, the incentives change.
The lesson from Addis Ababa's vehicle market, and from the repeated disputes over tax compliance, should therefore not simply be that businesses need to pay more tax.
They should. Nor should the lesson be that tax authorities need to become more aggressive. They need to become more capable. The distinction matters.
Ethiopia needs a revenue administration that can follow economic activity from import to sale, from supplier to retailer and from transaction to payment. It needs institutions that exchange information rather than merely exchange letters. It needs enforcement that is targeted by evidence and tempered by due process. And it needs rules that give legitimate businesses a clear path to compliance.
The ultimate test of tax reform will not be the number of shops inspected, vehicles impounded or taxpayers penalised.
It will be whether the state can make the taxable economy increasingly visible without making legitimate commerce unnecessarily difficult.
For too long, tax administration has often been forced to look for the transaction after it has disappeared from view.
The next phase of reform should make that unnecessary.
The state should not have to chase the economy. It should be able to see it.
