Fuel supply integrity: why the evidence trail matters

Fuel fraud follows price differences. A World Bank policy research paper by Masami Kojima (January 2016) explains that setting different prices for the same fuel depending on use, and large price gaps between fuels that can substitute for each other, create strong financial incentives for commercial malpractice. The paper names adulteration of diesel with kerosene as one of the most common forms. It notes that a considerable amount, "say 30 percent", can be added to automotive diesel without immediate detection, whereas even a 5 percent addition of kerosene to gasoline could lead to engine knocking. Of the 44 countries in the paper's price comparison, kerosene was cheaper than diesel in 31.

An earlier World Bank Viewpoint note by Kojima and Bacon (2001) describes smuggling, adulteration, mislabelling and short-weighting as widespread in many developing countries' fuel markets.

The same incentive exists wherever the same molecule is taxed or subsidised at two rates, or a cheaper product can pass for a dearer one, someone in the supply chain can profit from blending, diverting or misdeclaring it. Most of the parties who carry the cost, such as fleet operators, plant owners and end buyers, are not present when the product changes hands upstream.

Markets with routine monitoring still detect off-specification fuel. In the European Union, Article 8 of the Fuel Quality Directive (98/70/EC) requires member states to sample petrol and diesel sold for road transport each year and analyze it against the directive's limits. The sampling systems follow European Standard EN 14274. For 2024, the European Topic Center report published by the European Environment Agency on 26 June 2026 records 274 non-compliances for petrol and 31 for diesel across reporting countries. Belgium alone reported 84 for petrol and 14 for diesel, from 8,048 samples. Member states reported responses that included informing the competent authorities, opening investigations, imposing penalties and re-sampling.

In the UK, HMRC estimates the net diesel duty gap at 0.3 percent of theoretical diesel duty liability, or £40 million, in the 2024 to 2025 tax year, down from 5.7 percent in 2005 to 2006. Enforcement has relied in part on chemical markers in rebated fuel. When the UK and Irish governments announced a new marker in 2014, they described how launderers filter rebated fuel through chemicals or acids to remove the government marker, and how those residues remain in the fuel and damage fuel pumps in diesel cars.

These figures measure forecourt compliance and tax receipts. They say nothing about what happens to a given organization's fuel between the supplier's terminal and the point of use. A low national duty gap does not show that a depot's deliveries arrive on specification and in full.

Fuel changes custody several times before it is burned. It moves from the supplier's terminal to a road tanker, from the tanker into a depot or site tank, from the tank through a dispenser, and into a vehicle, generator or process. Each transfer is a point where volume can be short, product can be substituted or mixed, and water or contaminants can enter.

When an organization records only the invoice and the monthly consumption, an unexplained loss can be seen in aggregate but cannot be placed. A short delivery, product lost from a tank and dispensed volumes booked against the wrong asset all produce the same monthly variance, and only a record at each handover separates them. The same applies to quality. An engine failure blamed on fuel cannot be traced to a delivery unless a sample from that delivery was taken, sealed and kept.

In a dispute with a supplier or an insurer, the operator then has the invoice, the supplier's delivery note and the supplier's certificate, and little evidence of its own.

A workable evidence trail can start with the instruments already on site, provided each custody point produces consistent records that link to one another.

At delivery, record the delivery note and the supplier's certificate of quality for the batch, measured tank levels before and after discharge, the product temperature where volumes are temperature-corrected, and a retained sample with a seal number and a chain-of-custody record. In storage, record periodic tank readings reconciled against deliveries and dispensing, along with water checks and the results of any housekeeping. At dispensing, record volumes by asset and by authorized user.

The records must be joined. A laboratory result is of limited use unless it can be matched to a batch, a delivery and a tank, and a loss figure is of limited use unless it can be split by custody point. The national monitoring regimes offer a model for proportionate sampling. EN 14274 sizes the sampling program to the market and the statistical model in use, and an operator can size its own sampling plan to the volume and risk at each site.

When something does go wrong, the investigation record matters as much as the routine data. It should hold the samples tested, who handled them, the results and the decision taken, so that the case can be reviewed later by an auditor, an insurer or a court.

Procurement can write the evidence trail into supply contracts. That means a named fuel specification, a certificate of quality per batch, the buyer's right to sample at delivery and to retain samples for a stated period, agreed tolerances for volume differences, and the process and timetable for a dispute.

Operations can begin with reconciliation. Pick the sites with the largest throughput and the least explained variance, reconcile deliveries, tank readings and dispensing for a defined period, and record where the numbers do not close. The pattern of gaps shows where records are missing and where a loss is most likely to occur.

Finance and risk should agree how a fuel discrepancy is classified, investigated and closed, so that losses do not sit unexamined in a cost-of-sales line. THEFOCUS is developing FuelVeriq, a pre-launch product currently open to pilot enquiries, around this record-keeping problem. Whatever tooling is used, the controls above can start with the documents and procedures an operator already has.