The Subsidy, the Queue and the Price

On the morning of 30 May 2023 there were queues at filling stations across Nigeria, some of them a kilometre long.

This is worth pausing on, because at that hour almost nothing had actually happened. No instrument had been signed. No circular had issued. No regulation had been amended, no gazette published, no importer’s licence altered. What had occurred was a sentence in a speech at Eagle Square the previous day: subsidy is gone.

The queue did not form because the price had risen. At most stations it had not yet risen. The queue formed because several million people, independently and without consulting one another, reached the same conclusion at the same moment about what a sentence meant — and each of them acted on it before the others could.

The country answered before the state had finished speaking.

Everything that followed over the next three years — the prices, the reversals, the back-door arrangements, the refinery, the dollar-priced week in July 2026 — works out that first movement. The instrument was removed. What it had been holding down was still there, and it had to go somewhere.

The word carries a moral charge in Nigeria that obscures a fairly simple mechanism.

A fuel subsidy is not a gift, and it is not a payment to citizens. It is a purchase — the state buying a price. Petrol costs what it costs to land, refine, transport and dispense; the state chooses a lower number for the pump and then finds the money for the difference, whether by direct payment to importers, by an accounting device inside a national oil company, or by simply not being paid what it is owed. The difference exists whether or not anyone budgets for it.

Three consequences follow, and each of them shaped the Nigerian experience.

The first is that a subsidy is a quantity commitment disguised as a price commitment. The state promises a price but ends up owing on volume, and volume is decided by consumers rather than by the treasury. A subsidised price therefore has no ceiling on its cost. By early 2023 the monthly bill on the national oil company’s books had passed four hundred billion naira, and nobody had voted for that figure, because it was not a figure anyone chose.

The second is that a subsidy on a portable good is a subsidy to a region, not to a country. Petrol cheap in Nigeria and dear across a land border will cross it, and no enforcement changes that arithmetic — the same arithmetic that governs a currency premium. Some part of the national consumption figure was never Nigerian consumption, and nobody has ever been able to say how large that part was.

The third is the one that matters most for what came after. Because the subsidy paid a difference rather than a price, its cost rose whenever the naira weakened, and it rose without any decision being taken. The fiscal exposure was therefore tied to the exchange rate — and Nigeria’s exchange rate, as at May 2023, was itself an administered number carrying a premium of about sixty-three per cent. Two fixed prices were propping each other up, and each made the other more expensive to maintain.

A subsidy does not lower a cost. It moves it to someone who is not in the room.

The queue of 30 May 2023 is the clearest thing in this essay, and it demonstrates the sub-pillar of this cluster in a single morning.

Scarcity does not wait for prices. When a price is expected to rise and has not yet risen, the good is temporarily underpriced, and an underpriced good is rationed by whatever mechanism is available. That morning the available mechanism was time. Those who could spend six hours in a queue bought at the old price. Those who could not — anyone on a shift, anyone with a shop to open, anyone caring for a child — bought later at the new one.

This is the general law: when price stops doing the allocating, something else starts, and the something else is almost never fairer. A queue looks egalitarian because everyone stands in the same line. It is not egalitarian, because an hour is not worth the same to everyone in it. Rationing by time is a levy charged in inverse proportion to the value of the payer’s time, which means it falls hardest on people who are paid by the hour and lightest on people who can send someone else.

The same morning produced a second mechanism. Stations expecting to buy their next load higher had every reason not to sell the current one cheaply, and stock became unavailable — reported as hoarding, which is an accusation rather than an analysis. A trader who sells today at a price that will not replace tomorrow’s stock is not generous; he is liquidating.

Six instruments and events, in sequence. Only the first is a rule in the ordinary sense; the rest are the state, the market and the country answering it in turn.

This is the part of the subject where honest analysis is hardest, because both of the popular positions contain a real finding and neither contains the whole one.

The case against the subsidy is strong and it is distributional. Petrol consumption rises with income: a household that owns a car, runs a generator and travels by private vehicle consumes vastly more subsidised fuel than a household that does not. A universal price subsidy therefore transfers most of its value to the people who consume most, which means upward. Add the volumes that crossed borders, add the opacity of arrangements that were never appropriated, and the instrument was regressive, uncapped and unauditable at once. That is not a small case. It is close to a decisive one.

The case against removal is also strong, and it concerns incidence rather than intention. A poor household consumes little petrol directly and is exposed to its price through everything else — the fare, the food that arrived by road, the grinding machine, the small generator that is the only power the shop has. The benefit was concentrated at the top; the pain of withdrawal was general, immediate and unbuffered, and it arrived alongside a currency movement raising the price of everything imported. Both were consequences of policy, and they landed together on people with no capacity to adapt to either.

Both findings are true. The instrument was a bad instrument and its withdrawal was, for millions of households, a catastrophe — and no analysis that needs one of those to be false is worth reading.

The subsidy’s benefit was concentrated. Its removal’s pain was general. That asymmetry is the entire politics of the thing.

Holding both does not produce paralysis. It produces a question about sequencing and compensation. The savings were real and the promise attached to them was explicit — that the money would go into infrastructure and services. Whether it did, and at what scale, determines which of the two findings above eventually dominates. That assessment requires evidence this cluster has not yet gathered.

Three years on, the question is still being asked in public, and the reason it can be asked at all tells us something about the instrument.

A large receivable said to sit on the national oil company’s books has been raised in public debate as evidence that the obligation never truly closed. The figure is contested and this essay does not adopt it; what matters is structural rather than arithmetical. When a subsidy is abolished by declaration rather than by an instrument that also disposes of the arrears, the accumulated difference does not vanish. It sits somewhere — as a receivable, an arrear, a non-payment, an adjustment — and it will be settled eventually by someone, most probably by the public, through a mechanism nobody will get to vote on.

The honest answer has three parts. The administered pump price is genuinely gone: it now moves with crude and with the naira, and it moved twice in a single week in July 2026 with no state instrument involved. The fiscal exposure to a subsidy is largely gone too, which was the reform’s central purpose. But the country’s exposure to the price of imported energy is not gone at all — it has moved from the fiscal accounts, where it was hidden and unbounded, to the pump, where it is visible and weekly.

Three consequences follow for anyone reading this instrument, whether as a household or as a business.

1. Treat the pump price as an exchange-rate instrument. Since deregulation it responds to crude and to the naira, and the July 2026 episode showed that the transmission is now direct enough to move retail prices twice in a week.
2. Distinguish abolition from disposal. An obligation declared over is not an obligation settled, and the difference between the two is where fiscal surprises are stored.
3. Watch the number of sellers, not the price. A falling price under one dominant supplier is a decision; a falling price under several is a market. The February 2026 reduction was welcome, and it is not yet evidence of the second.

Compared with what? Against continuing an uncapped, regressive and unappropriated instrument indefinitely, removal was defensible and probably necessary. Against a phased withdrawal announced in advance, sequenced apart from the currency float, and accompanied by transfers that reached households before the price moved rather than after, the actual removal looks considerably worse. The relevant comparison was never subsidy or no subsidy. It was this removal or a better-sequenced one — and that is the comparison a single sentence in an inaugural address makes it impossible to hold.

Return to the queue.

It was there before the price moved, before any instrument existed, before anyone could have read a document. Several million Nigerians heard a sentence, understood exactly what it implied for the cost of their week, and acted within hours — which is to say that the country’s economic literacy was never the problem. People understood the instrument perfectly. They had been living inside it for their entire lives.

What they did not have was any way to act on that understanding except by standing in line. That is what it means to be governed by an instrument you cannot see, cannot cost, and were never asked about: you find out what it was worth on the morning it is withdrawn.

The subsidy was never a gift. It was a bill that had not yet been addressed to anyone.