Editorial Article

On 14 June 2026 Honey Labels Start Asserting Percentages. No Published Method Can Prove One.

Updated November 21, 2026

On 14 June 2026 a jar of honey sold in the European Union has to carry a number it has never carried before. If the honey is a blend, the label must state the countries of origin in descending order of their share by weight, and it must state the percentage each of those countries represents. The permitted tolerance is five percentage points on each share.

That is a quantitative obligation about provenance. And there is no published analytical method that can discharge it.

This is not a small drafting gap, and it is not the sort of thing that gets fixed quietly. It determines how an entire category of food testing behaves for the next three years, and it explains why the authenticity testing market grows at 7.67 per cent a year while capturing almost no price.

What the Directive actually says

Directive (EU) 2024/1438 of 14 May 2024 amends four marketing directives at once — honey, fruit juices, jams and marmalades, and certain preserved milks — which is why it is known in the trade as the Breakfast Directive. It was published in the Official Journal on 24 May 2024, had to be transposed into national law by 14 December 2025, and its rules apply from 14 June 2026. France transposed by Decree No. 2026-312 of 24 April 2026, in force on the same date.

For honey, the new origin provision is precise. The country or countries where the honey was harvested must be indicated on the label. Where honey originates in more than one country, the countries must appear in the principal field of vision, in descending order of their share by weight, together with the percentage each represents. The tolerance is 5 per cent for each individual share within the blend. The figures become mandatory particulars under Regulation (EU) No 1169/2011, which attaches the general labelling enforcement machinery to them.

Two reliefs were written in. Where a blend draws on more than four countries and the four largest shares together exceed 50 per cent, Member States may allow percentages for those four only, with the remaining countries listed in descending order and unquantified. And where a package contains less than 30 grammes of honey, the country names may be replaced by two-letter ISO 3166-1 codes.

Then the clause that matters most, and it is easy to read past: the tolerance is calculated on the basis of the operator’s traceability documentation.

That is where the percentage comes from. Not from a laboratory. From the paperwork.

What analysis can and cannot do

It is worth being exact about the current state of the science, because the gap is narrower than “nothing works” and wider than “it’s fine”.

Detection of added sugar works, and has had to keep working. For two decades the workhorse method was stable carbon isotope ratio analysis by elemental analyser isotope-ratio mass spectrometry, AOAC method 991.41, which exploits the fact that maize and sugarcane are C4 plants and carry a carbon signature that honey, built from C3 nectar, does not. The Joint Research Centre put that method against 320 imported consignments and reported that it was not effective in detecting honey suspicious of being adulterated — because the syrups had moved to rice, wheat and sugar beet, all C3 plants. The adulterant changed its isotope chemistry to defeat the reference method. What replaced the single method was a panel: EA-IRMS and liquid chromatography IRMS, high-performance anion-exchange chromatography with pulsed amperometric detection, and proton nuclear magnetic resonance, used together.

Origin consistency can be assessed. Proton NMR profiling compares a sample against a database of characterised references. Bruker’s FoodScreener honey method, run at 400 MHz, holds roughly 28,500 reference samples spanning more than 50 countries, 100 monofloral types, seasons, production years, blends and production methods, and reports on exogenous sugars, country of origin, botanical variety, composition and freshness.

Botanical variety can be assessed in the same run, and separately by pollen DNA. New Zealand’s Ministry for Primary Industries defines monofloral manuka in five attributes, one of which is a DNA marker from manuka pollen; a United Kingdom laboratory prices that marker alone at GBP 246 per sample.

What none of it does is quantify a share of a blend. A database can tell you a sample is consistent with an origin, or inconsistent with it. It cannot tell you that 34 per cent of the jar came from Vietnam and 21 per cent from Ukraine. Establishing a proportion to within five percentage points across an unknown number of components is a different analytical problem from establishing consistency, and no catalogue method does it.

So the label asserts a percentage that is derived from records and cannot, today, be tested against a measurement. That is why this market’s product is a suspicion rather than a verdict, and it is why the European Commission states in its own enforcement findings that current methods are not developed enough for a simple test to be sufficient and that proper follow-up is necessary.

The enforcement record, for scale

The Commission has run the exercise. In the coordinated control plan on imported honey, 320 consignments originating in 20 exporting countries were sampled by 15 Member States plus Switzerland and Norway between October 2021 and February 2022 and analysed centrally. 147 of them — 46 per cent — were suspicious of non-compliance with the prohibition on added ingredients, against a 14 per cent rate in a comparable exercise in 2015–17. The absolute count was highest for China at 66 of 89 consignments; the relative rate was highest for Turkey at 14 of 15 and for honey arriving from the United Kingdom at 10 of 10.

Of the operators behind those consignments, 57 per cent had exported suspicious goods and 66 had imported at least one. Forty-four EU operators were investigated and seven were sanctioned. One in six.

And the forensics produced a detail with no equivalent in any other food safety category: the recorded modus operandi included analysis in accredited laboratories to adapt honey and sugar blends so as to elude possible detection by clients and official authorities before import operations. The offender bought the test in order to defeat the test.

Put those three findings together and the regulatory logic becomes clear. Detection is probabilistic, suspicion is not proof, and the party being tested has access to the same analytical capability as the party doing the testing. A percentage on a label is therefore, in 2026, a documentation claim enforced by audit, not a measurement enforced by analysis.

The timeline that decides the market

Date What happens Analytical consequence
14 May 2024 Directive (EU) 2024/1438 adopted —
24 May 2024 Published in the Official Journal —
14 December 2025 Deadline for transposition into national law National enforcement structures align
14 June 2026 Rules apply: origin declared in descending share, with percentages and a 5% tolerance Origin becomes a falsifiable quantity; no method exists to measure it
14 June 2029 Deadline for delegated acts on composition criteria, pollen integrity, harvest location and Union-wide traceability Potential arrival of numerical limits — the thing that creates price

The 2029 deadline is the one to watch, and it is worth being specific about what is coming. Article 4a empowers the Commission to adopt delegated acts on: the criterion of “mainly” as regards floral or vegetable origin; composition criteria to ensure honey has not been heated or treated so as to destroy or significantly deactivate natural enzymes, by reference to the invertase index; criteria to ensure and verify that pollen is not removed from honey and that absolute pollen content and pollen spectrum are not modified; minimum pollen content in baker’s honey; and the methods and criteria to determine where honey was harvested, plus Union-wide traceability requirements.

That list is a specification for the next generation of this market. An invertase index and a pollen-content threshold are numbers. Enforcement against a number requires a method and a reference material. The Commission must run feasibility studies first, and the study on harvest location must analyse digital solutions including a unique identifier code.

If those acts arrive as numerical limits, the category converts from an investigative service into a compliance product — and the price behaviour changes with it. In the Pheonix Research cluster of five food testing reports, the categories where a threshold is published capture between a third and a half of their growth in price. The category where no threshold exists gives price back.

The American answer is paperwork, and that is the point

The United States has looked at the same problem and bought something different. The FDA’s Food Traceability Final Rule under Section 204(d) of the Food Safety Modernization Act requires firms handling foods on the Food Traceability List to keep records of Key Data Elements at Critical Tracking Events and produce them within 24 hours. The original compliance date was 20 January 2026; the FDA proposed a 30-month extension to 20 July 2028, and the Continuing Appropriations, Agriculture, Legislative Branch, Military Construction and Veterans Affairs, and Extensions Act of 2026 directed the agency not to enforce the rule before that date, a directive the FDA has said it intends to follow. The agency also convened a public meeting on lot-level traceability on 6 November 2026.

FSMA 204 answers “where did this lot come from and where did it go”. It does not answer “is this what it says it is”. Both are legitimate strategies and they buy different things: the EU is buying determinations, the United States is buying provenance. The commercial consequence is that growth in authenticity testing is EU-weighted, which is why the Pheonix Research model states that reading qualitatively rather than pretending to a regional split it does not carry.

What this means if you sell food

Three practical consequences follow from the gap between the obligation and the method.

Your traceability record is now your evidence. The percentage on the jar is calculated from your documentation and the five-point tolerance is calculated from it too. If the record is thin, the label is exposed regardless of what the honey actually is.

Analysis is a defence, not a certificate. A consistent NMR profile corroborates a declared origin. It does not prove a share. Buying more assays does not close a records gap, which is why the advisory and provenance layer of this market grows faster than the analytical layer.

The price of certainty is about to change. Until the 2029 delegated acts, authenticity testing remains a discretionary purchase priced by how many claims a label asserts. After them, it may become a required purchase priced by a limit — and the buyers who build the records now are the ones who will find the transition cheap.

This article draws on the Pheonix Research report Global Food Authenticity & Adulteration Testing Market, 2026–2033, Directive (EU) 2024/1438 and the European Commission’s EUR-Lex summary of its application dates, the Commission’s coordinated control plan findings for honey (2021–2022), the US FDA’s Food Traceability Rule compliance page, and the JRC’s certified reference materials for fish species identification. Related reading in this cluster: olive oil, where an established chemical standard supports confirmatory testing, and plant hygiene monitoring, where an unpublished threshold has the opposite effect on price.

Read the companion piece: The Same Honey, Five Times the Invoice — on why a food authenticity determination is priced by how many things a label asserts rather than by what the laboratory does.

Frequently Asked Questions

Can a laboratory test prove the percentages on a honey label?

No. NMR profiling and isotope-ratio mass spectrometry can establish whether a sample is consistent with a declared origin, and can detect added sugar syrups, but they do not quantify the proportion of each origin in a blend to within five percentage points. The Directive itself calculates the tolerance from the operator's traceability documentation rather than from analysis.

What exactly changes in EU honey labelling in June 2026?

Directive (EU) 2024/1438 was adopted on 14 May 2024, had to be transposed by 14 December 2025, and its rules apply from 14 June 2026. A blend must declare the countries of origin in descending order of share by weight, with the percentage each represents, subject to a 5 per cent tolerance per share.

When do the composition and pollen rules arrive?

By 14 June 2029 the Commission must adopt delegated acts on the criterion of 'mainly' for floral or vegetable origin, composition criteria by reference to the invertase index, criteria on pollen removal and pollen content, minimum pollen content in baker's honey, and methods and criteria to determine where honey was harvested plus Union-wide traceability. Those are the provisions that could turn authenticity into a compliance market with numerical limits.

Does the US require the same testing?

No. The FDA's Food Traceability Final Rule under FSMA Section 204(d) requires records of Key Data Elements at Critical Tracking Events, with a compliance date moved to 20 July 2028. It is a documentation rule about provenance, not an analytical rule about composition or species. That is one reason growth in authenticity testing is concentrated in the EU.