// analysis
Emergency oil reserves: from the market to the pump
From emergency stocks to a fill-up: prices, exchange rates, taxes and margins. French data show why consumer savings need careful measurement.
Investigation: “Emergency fuel reserves and the delivery test” · Part 6 of 7 · Sources checked through 3 October 2026.
Fifty litres of diesel would cost €118.53 at France’s average pump price on 25 September 2026. The energy ministry’s DGEC recorded €2.3706 a litre, compared with €2.3834 a week earlier. On that notional fill-up, the saving was 64 cents. These are weighted averages for mainland France excluding Corsica; an individual station can charge a different price. S02 S03
On 2 October, the G7 announced a coordinated release of 100 million barrels over four months, with a substantial diesel component brought forward into the first twenty days. The statement places the operation within the implementation of commitments made in March, taking account of commitments already fulfilled. For drivers, the practical question is how much of that intervention will reach the forecourt price board. S01
The dates establish the starting point. The decline between 18 and 25 September happened before the announcement. It provides a useful example of price pass-through, rather than evidence of the G7 operation’s effect. Following the money to the customer requires matching the product, currency and observation period, then examining distribution costs and taxes.
Two clocks start running
A reserve release can influence prices before the first truck leaves a terminal. Buyers reassess the supply they expect to obtain, sellers reconsider their opportunities, and inventory holders weigh the value of keeping barrels for later. A credible delivery tomorrow can change a bid today. Lutz Kilian and Xiaoqing Zhou’s work on the US Strategic Petroleum Reserve explicitly considers the relationship between public intervention, expectations and private storage. S11
That channel depends on the announcement’s credibility. Product specifications, location, access conditions and timing turn an aggregate volume into a commercial prospect. A promise that will be difficult to execute gives buyers less reason to lower their bids. Part of the effect may also precede the official decision when the market already expects the operation. This describes a possible mechanism; it is not an estimate of the market’s response on 2 October. S11
The other clock measures operations: awards, lifting, any processing still required, and delivery. Earlier instalments followed those physical constraints. They return as price pressure when buyers compete for insufficient supplies in the same place. The US Energy Information Administration describes how low inventories can prompt wholesalers to bid more aggressively, while limited transport capacity sustains regional price premiums. S12 S13
The contents of a reserve determine the remaining journey. Crude goes through a refinery. Finished diesel that meets the required specification can enter distribution more directly. The first relieves the plant’s feedstock needs; the second supplies the product buyers are seeking. Equipment availability and processing costs lie between them. That distinction helps explain why crude and diesel prices can respond differently. S04 S12
The economic reach of a release therefore extends beyond the movement of its own molecules. A station may benefit from a less strained wholesale market while continuing to buy from its usual supplier. Conversely, taking delivery of a reserve cargo establishes a supply transaction, but leaves its purchase price, resale conditions and local competition to be examined. A commercial benefit and a consumer saving occur at different points in the chain.
Brent changes currency on its way to the pump
The barrel on a financial screen is often Brent. A distributor buys a refined fuel with a specification, a delivery point and contractual terms. DGEC distinguishes regional product quotations, refining costs, and transport and distribution costs. Diesel can remain scarce while crude becomes cheaper. S04
Currency adds another movement. In the bulletin dated 25 September, dated Brent falls from $123.83 to $117.81 a barrel between the two reported weeks. Calculated from those rounded levels, that is a decline of roughly 4.86%. In euros, the published values move from €107.31 to €103.01, a 4.01% decline. Exchange-rate movements soften the reduction in the currency relevant to French buyers. This is dated Brent, distinct from the futures contract also shown in the bulletin. S02
The same document puts refined diesel at $1,466 a metric tonne, down from $1,532, a calculated reduction of 4.31%. The tax-inclusive pump price falls 0.54% between the two Fridays. Those differences deserve investigation. Interpreting them requires identifying what each series measures: crude or finished product, dollars or euros, a quotation average or a retail snapshot, a price before or after taxes. S02 S03
Dividing 0.54 by 4.31 would produce a misleading “pass-through rate”. Friday’s retail price may still reflect earlier procurement costs, while the wholesale quotation has just moved. Taxes introduce another difference in the denominator. A meaningful estimate follows the sequence of adjustments after a change in the relevant input cost, allowing for other changes in the meantime.
What moves inside a litre’s tax bill
For 25 September, DGEC reports a diesel price excluding all taxes of 136.80 euro cents a litre. The 2026 energy-tax guide lists a reference B7 diesel excise duty of €60.75 per hectolitre, equivalent to 60.75 cents a litre. Mainland France’s 20% VAT applies to the combined pre-tax price and excise duty. S02 S05
Using the published figures reproduces the bulletin’s tax-inclusive price: 136.80 cents before taxes, 60.75 cents of duty and 39.51 cents of VAT, for a total of 237.06 cents a litre. The pre-tax component includes fuel procurement and the economics of distribution. Labelling all of it “crude oil” would remove an important part of the supply chain from the calculation. S02 S04 S05
Now consider an explicitly hypothetical example. The retail price excluding all taxes falls by ten cents a litre, with tax rules unchanged. Excise stays constant. VAT falls by two cents, or 20% of the reduction in its taxable base. The customer then saves twelve cents a litre, or six euros on fifty litres. This arithmetic assumes the full ten-cent reduction reaches the pre-tax retail price; it predicts no change in the market. S05
On 22 September 2026, the government proposed directing any additional fuel-tax receipts towards support measures through an amendment to the 2027 finance bill. The announcement concerns how revenue would be used; its effect would depend on the legislation adopted and the support arrangements. The preceding scenario holds existing tax rates constant. S17
Keeping cents separate from percentages avoids a common mistake. A fixed duty dampens the percentage movement in the final price because it enlarges the initial amount used as the denominator. VAT increases the movement in cents. Saying that “taxes absorb the fall” without specifying the unit conflates these two effects.
A lower wholesale quotation may be partly offset by another cost before it reaches the pre-tax retail price. Procurement invoices and margins matter at that stage. Tax arithmetic describes what happens next. It also separates market movements from direct price interventions. Mainland continental France’s 30-cent-per-litre rebate including VAT, introduced on 1 September 2022, was a distinct event that an analysis of that period must identify. S07 S16
Adjustment lags measured by the Banque de France
Research published by the Banque de France on 14 October 2021 provides an empirical reference. Following a 1% rise in the euro price of refined diesel at Rotterdam, the authors report a 0.45% increase in French stations’ pre-tax diesel price after one week, 0.72% after fourteen working days, and 0.75% in the long run. Full adjustment takes roughly twenty working days in their analysis. S06
Those figures describe a gradual process. The first represents 60% of the estimated final response; the second, 96%. Fuel bought between those dates is priced in a market where earlier changes are still working through. A recent reversal in wholesale quotations can temporarily coexist with retail increases reflecting the previous upswing. That possibility follows from adjustment lags; it remains something to test in each episode.
In the case they study, the authors find similar responses to rising and falling costs. That is relevant evidence against assuming that distributors systematically withhold price cuts. It leaves other periods and individual practices open to investigation: the Banque de France itself notes that the wider literature does not offer a uniform conclusion. S06
The date deserves as much attention as the coefficients. Although the page carries an update dated 24 September 2026, the post and the estimates presented were published in 2021. The exact end of the sample is not specified. The study provides a method and a historical benchmark. Predicting how French stations will adjust over twenty working days in October 2026 would require a new estimate. S06
The 0.75% figure also needs its proper denominator. The pre-tax retail price includes more than the raw material. A 1% increase in one component produces a smaller percentage movement in the whole. Reading the difference as “25% of the increase retained” would confuse percentages calculated on different starting amounts.
The margin in the chart and the costs behind it
To monitor distribution, the French finance ministry estimates the gap between the pre-tax retail price and a refined-product benchmark. This gross transport and distribution spread must cover storage, delivery, station operations, some regulatory costs and the commercial return. Reporting it as net profit would treat the money used to pay those expenses as earnings. S07 S08
The comparison’s timing matters. The methodology published on 6 May 2026 uses quotations from five previous working days, J−2 through J−6. That lag is intended to approximate logistics and inventory turnover. The detailed document combines several supply regions, weighted according to the networks’ sourcing. S07 S08
Suppose a quotation falls sharply today. Comparing it with today’s station price mechanically produces a larger gap than comparing the station price with earlier quotations. The size of that difference depends on the movement in the market and the lag chosen. Establishing a change in profitability would still require actual procurement prices, operating costs and sales volumes. The official methodology acknowledges these limitations and applies a common convention to businesses with different contracts. S08
DGCCRF’s monitoring is an economic estimate. It uses 2024 volume weights and a common quotation lag that cannot reproduce each network’s procurement terms. Stations without available volume data are excluded from the weighted calculation. The indicator therefore does not measure distributors’ net profit. S08
That leaves plenty of room for independent scrutiny. In a publication dated 11 March 2026, updated on 20 March, the consumer association CLCV reports a rise in its diesel gross-distribution-margin indicator in early March relative to its 2025 average. It is a signal from a consumer advocate over a defined comparison window. Matching the scope and then examining accounting evidence would be necessary before reaching a conclusion about realised profit. S09
A long average can hide a brief spike; a brief spike can coexist with an average that has returned to a familiar level. Tracking rises and falls using the same product, benchmark and lags makes it possible to look for asymmetry rather than assume it. A persistent discrepancy that survives those checks is a stronger lead than a screenshot taken on a day when Brent falls.
The price without the intervention
An intervention can moderate a rise without making the observed price fall. Equally, a price may decline mainly because demand weakens or another source of supply returns. Attributing a saving to a stock release requires a counterfactual: an estimate of the price that would have prevailed in the same circumstances without the operation. Kilian and Zhou place that comparison at the centre of their approach. S11
The US experience in 2022 illustrates the importance of assumptions. On 26 July, the Treasury estimated that US and partner IEA releases had reduced gasoline prices by 17 to 42 US cents a gallon. The calculation assumes full pass-through and uses two values for demand’s sensitivity to price. The authors themselves warn that refining constraints could limit transmission. The range reflects the model’s sensitivity to those choices, rather than a statistical confidence interval. S10
This assessment came from the administration responsible for the intervention. It offers a method with identifiable parameters and limitations, rather than a conversion rule from millions of barrels to savings at a French filling station. Switching product, country or crisis changes the conditions of the calculation.
Private inventories add another complication. A barrel leaving a public reserve may help rebuild an operator’s own stock. Depending on the circumstances, it reduces exposure to a future disruption or adds to the supply immediately offered for resale. Consumption and public-reserve withdrawals can therefore follow different schedules. Explicit treatment of that storage demand is one contribution of Kilian and Zhou’s analysis. S11
The French series verified here ends on 25 September, before the G7 announcement. It establishes the starting conditions and the small reduction on the final Friday. Actual subsequent deliveries, changes in procurement costs and the programme’s own effect on prices remain to be established. Extending a line beyond its observations would imply evidence that this collection does not provide.
Follow the same stations, then examine the contracts
An evaluation would begin by following the same stations before and after the operations, fuel by fuel. France’s public portal provides timestamped prices and station information, with archives extending back to 2007. It offers a foundation for that work. Daily sales volumes and purchase invoices needed for a complete margin analysis are outside the fields described. No station-level econometric panel has been estimated for this article. S14
Stations that run out of fuel would also need attention. A price still displayed for unavailable diesel says little about the cost to a driver who must buy elsewhere. The changing composition of observed outlets, customer movements and weighting choices influence an average. DGEC uses volume and market-share weights. An unweighted average of public station records would measure something different. S03 S14
The next stage would match prices with the relevant wholesale quotations, exchange rates, logistics costs and tax or commercial changes. Isolating a reserve release remains difficult because it happens during a disruption and can affect markets across several countries. An apparently untreated comparison region may benefit from the same easing in wholesale conditions. That research design would require procurement data and a new estimate; no results from such an evaluation are presented here.
The financial route is now identifiable. Credible additional supply can soften bidding pressure; procurement-cost movements work gradually into retail prices; taxes then translate that movement into the amount paid. Historical adjustment estimates and the litre’s price decomposition provide a more precise way to read the next observations. The saving attributable to the 2 October announcement remains to be established.
The final instalment will follow the operation in reverse: replenishing the reserves. Purchase dates, prices and repayment commitments will determine how the cost of emergency supply is distributed over time.
Further reading
- Part 1: the journey from storage to diesel.
- Part 2: who owns France’s emergency oil?.
- Part 3: the distance to the pump.
- Part 4: the refinery sets the pace.
- Part 5: the contracts behind the release.
- Strategic oil reserves buy time.
- Read the oil market.
Sources
- G7 / European Council : G7 Leaders’ Statement on global energy security and market stability. 2026-10-02. Announcement on 2 October 2026; four-month programme
- DGEC : Oil quotations, prices and margins in France and the European Union, bulletin dated 25 September 2026. Document accessed 3 October 2026. Bulletin dated 25 September 2026; retail observations 3 July–25 September; quotation weeks labelled 18 and 25 September
- DGEC : Method for calculating average petroleum-product prices. Document accessed 3 October 2026. Methodology updated 6 January 2025
- Ministère de la Transition écologique / DGEC : Petroleum-product prices. 2019-07-05. Explanatory page and data portal accessed 3 October 2026
- Ministère de la Transition écologique : 2026 guide to energy taxation. Document accessed 3 October 2026. 2026 guide; annex as at 1 August 2026
- Banque de France ; Erwan Gautier, Magali Marx, Paul Vertier : How do oil prices pass through to fuel prices?. 2021-10-14. Results published in 2021; sample end date not specified in the post
- Ministère de l’Économie : Government monitoring of gross fuel-distribution spreads. 2026-05-06. Press release 6 May 2026; gross spread methodology
- DGCCRF : Mainland fuel: average gross transport and distribution spreads. Document accessed 3 October 2026. Method: quotations lagged by J−2 through J−6 and 2024 volume weights. Recent numerical results are not reproduced.
- CLCV : Fuel prices: CLCV reports higher refining and distribution spreads. 2026-03-11. Early March 2026 compared with the 2025 average
- U.S. Department of the Treasury ; Benjamin Harris, Catherine Wolfram : The Price Impact of the Strategic Petroleum Reserve Release. 2022-07-26. Assessment of 2022 US and partner releases; US gasoline
- Federal Reserve Bank of Dallas ; Lutz Kilian, Xiaoqing Zhou : Does Drawing Down the U.S. Strategic Petroleum Reserve Help Stabilize Oil Prices? Working Paper 1916. Posted 20 December 2019; version dated 19 December 2019. Historical global oil-market modelling; version dated 19 December 2019
- U.S. Energy Information Administration : Factors affecting diesel prices. Document accessed 3 October 2026. Economic mechanisms; page accessed 3 October 2026
- U.S. Energy Information Administration : Gasoline price fluctuations. Document accessed 3 October 2026. Explanatory discussion of inventories and price fluctuations
- Ministère de l’Économie : Public fuel-price data. Document accessed 3 October 2026. Archives from 2007 and price feeds; portal accessed 3 October 2026
- Commission européenne : Weekly Oil Bulletin. Document accessed 3 October 2026. Bulletin shown dated 1 October 2026; price tables labelled 28 September, corresponding to Friday 25 September for France; French excise applicable from 1 August 2026
- Gouvernement français : France’s fuel rebate from 1 September 2022. 2022-08-26. Fuel rebate in mainland France from 1 September 2022
- French government : Fuel taxes: the government’s proposed revenue rule. 2026-09-23. 23 September explanation of the proposal announced on 22 September 2026; intended amendment to the 2027 finance bill
Scope and method
Price series retain their observation dates. The Banque de France coefficients are historical estimates published in 2021; the tax scenario is hypothetical. DGCCRF’s monitoring is used here to define a gross spread and its limitations, without estimating net profit or reproducing its recent numerical results. No new causal effect of a reserve release is estimated.
The five CSV files reproduce the rounded values in the DGEC bulletin dated 25 September and the calculations explained in the figures. The European bulletin published on 1 October was also downloaded. Its tables, labelled 28 September, report France’s prices from the previous Friday, 25 September. They corroborate the last observation in the series and add no price observed after the G7 announcement. S02 S03 S15
DGEC’s tables dated 25 September were checked in extracted text and rendered pages. Calculations use published rounded inputs. The evidence includes no individual procurement contracts, operator interviews or econometric station panel.
Reproducible inputs: price decomposition, hypothetical tax scenario, 2021 Banque de France estimates, thirteen Friday observations, September changes.
This analysis is not investment advice.
// cite this analysis
l0g, “Emergency oil reserves: from the market to the pump”, l0g.fr, published October 03, 2026, updated October 03, 2026, https://l0g.fr/en/analysis/emergency-oil-reserves-pass-through-pump-prices/
$ cd ../analysis