Cheap Oil Is a Loaded Spring

At the beginning of 2026, the maximum Nairobi price for a litre of super petrol was KSh 226.96. A month later, the posted ceiling was KSh 178.28. That was a drop of KSh 48.68 per litre—about 21 per cent—between two official pricing periods. The relief was real. So was the trap in treating relief as a forecast.

Oil prices can remain quiet long enough for households and governments to build budgets around them. Then a war, an export interruption or a shipping bottleneck changes the price of the next barrel. A calm month is evidence about the current market; by itself, it is no guarantee about the next one.

A price board is a lagging signal

Kenya’s Energy and Petroleum Regulatory Authority (EPRA) sets maximum retail prices monthly under a regulated formula. Its schedule for 15 January to 14 February 2026 lists Nairobi super petrol at KSh 226.96 a litre. The next schedule, in force from 15 February through 14 March, reduced it to KSh 178.28. These are regulated maximum prices for distinct periods, not a daily quote or a measure of what every pump charged.

The February price reflected earlier costs, taxes, transport and margins. It did not mean a new barrel had become permanently cheap. EPRA’s later March release reported that the average landed cost of imported super petrol rose from US$582.11 per cubic metre in February to US$823.87 in March, a 41.53 per cent increase. The equivalent landed cost for diesel increased 68.72 per cent, from US$636.45 to US$1,073.20 per cubic metre. By April, EPRA’s price schedule listed Nairobi super petrol’s landed cost at KSh 107.23 per litre. A separate government tax change then reduced the April pump price by KSh 9.37; the final retail figure was not a direct readout of crude alone.

That lag matters. A world-market shock can affect the cost of imports before it appears in the regulated pump price. Conversely, the price at a forecourt reflects more than the latest Brent quote: product type, shipping, refining, exchange rates, taxes, distribution, storage, margins and any stabilization measures all intervene. A household sees one number, but the number has a supply chain and a policy calendar behind it.

EPRA’s timing rules help explain the delay: its import-cost calculation uses a defined prior-month window, then it publishes a maximum price for the next period. A cargo bought or priced during a calm window can therefore keep the next schedule low even as a new disruption raises current futures prices. Later, a more expensive cargo can enter the calculation after the headline market has already retreated. The authority’s formula page describes the monthly review and cost components. This is a lag, not evidence that the market ignored the shock or that each household paid the same amount.

From quiet market to supply shock

In February, the global benchmark looked comparatively subdued. The Federal Reserve Bank of St. Louis’ FRED series, based on IMF primary-commodity data, puts the monthly global Brent benchmark at about US$70.89 a barrel. On 28 February, military action began in the Middle East. The U.S. Energy Information Administration later described a de facto closure of the Strait of Hormuz and reported that Brent front-month futures began the year around US$61 a barrel and ended the first quarter at US$118. Its first-quarter account says prices rose sharply after the military action and subsequent disruption.

The distinction between price measures is important. The US$70.89 figure is a monthly benchmark average; US$118 was the quarter-end price for a front-month futures contract. Neither is Kenya’s landed cost or the pump price. The EIA reported that Brent futures later ranged from US$72 to US$118 during the second quarter, with the high on 29 April, as flows through Hormuz remained disrupted. A market can retreat from a peak and still leave importers facing volatile, elevated costs.

The episode does not prove that a war or closure was predictable from February’s data. That is precisely the point: the price series described the calm regime before it changed. The International Energy Agency’s March report called the conflict the largest supply disruption in the history of the global oil market and described tanker flows through Hormuz as near halted. Those are retrospective institutional assessments, not evidence that every future route disruption will have the same scale or duration.

What the simple probability does—and does not—say

Suppose, for illustration only, that a “break” has a constant 10 per cent chance in each month and that the months are independent. The chance of at least one break in three months is:

1 − (1 − 0.10)^3 = 0.271, or about 27 per cent.

The calculation does not estimate the real chance of an oil shock. It assumes the very things that actual markets often violate: a stable monthly probability and independence. Wars, policy decisions, inventories and shipping constraints can persist across months, making events clustered rather than separate coin flips. The toy model simply illustrates why low monthly risk does not equal zero risk across a longer period.

It also does not justify assuming that each quiet month increases the probability of the next shock. A spring metaphor can mislead if it suggests pressure must mechanically build until it snaps. Markets are not clocks winding down. The useful distinction is between observed conditions and exposure to scenarios that have not happened yet.

The bill reaches farther than fuel

When imported fuel costs rise, the first effect is visible in transport and energy. The next effects travel through freight, farm inputs, food distribution and the operating costs of businesses. A matatu operator may face higher diesel costs; a trader may pay more to move stock; a clinic may pay more to run a generator or transport supplies. Each cost reaches a household differently, depending on where it lives, how it earns, and whether a service provider absorbs or passes through the increase.

The February KNBS inflation release already showed why an average can conceal strain: headline annual inflation was 4.3 per cent, but food and non-alcoholic beverages rose 7.3 per cent, while core inflation was 2.1 per cent. KNBS’s report records these as distinct measures. A household that spends much of its income on food may feel pressure even when the headline is moderate. Later fuel increases can add to the squeeze, but the sequence does not show that every food price moved because of oil alone. Weather, harvests, storage, currency and market structure also matter.

There is a countercase to the image of a system with no buffers. Oil companies hold inventories; governments can adjust taxes or use stabilization mechanisms; importers can diversify suppliers and shipping routes; households and firms can reduce or shift consumption. Kenya’s April schedule itself shows that the final pump price can be altered by policy: after a tax change, EPRA recalculated maximum prices. These buffers can soften an abrupt increase, although they distribute costs rather than erase them. Lower taxes mean less public revenue; subsidies require financing; inventories cost money and cannot cover every disruption indefinitely.

For a clinic, a practical resilience question is not simply “what will oil cost?” It is which essential activities depend on transport or backup power, how long fuel stocks last, and which purchases could be delayed without compromising care. For a household, flexibility is often much smaller: fare, food and medicine compete for the same cash. Policy buffers and personal buffers are not equally available to everyone.

Forecasts belong beside their dates

On 16 October 2026, the series’ forecast register assigned a subjective 65 per cent probability to Brent front-month settling above US$90 on 15 January 2027, and 25 per cent to Brent settling above US$120 on at least one day before that date. These are dated, uncalibrated judgments, not measurements of current market odds or financial advice. Their value is that each specifies an event and a resolution date, so the result can later be compared with the stated probability.

For a probability p and a yes-or-no outcome y, the Brier score is (p − y)^2: if a 65 per cent forecast resolves yes, its score is 0.1225; if it resolves no, the score is 0.4225. Lower is better, but one event cannot establish whether a forecaster is well calibrated. A record across many forecasts is needed, and even then the questions must be defined consistently.

The larger lesson is narrower than “cheap oil is dangerous.” A low price can be a welcome reprieve and still be a weak guide to the next shock. Distinguish monthly averages from futures settlements, global benchmarks from landed costs, and regulated pump prices from the cost of a particular household’s energy. Watch the route, the inventory, the tax and the lag between them. Calm is a condition. It is not a guarantee.

A Shell gas station in Nairobi beside the A104, photographed 11 July 2017. Historical file photo of a local retail fuel site; it does not show the 2026 EPRA price schedule or current pump prices. Credit: flightlog, via Wikimedia Commons. Image source · CC BY 2.0 Generic