What Testing Reveals, and What It Cannot, About Food Fraud in Global Supply Chains
An evidence-led analysis of supply-chain vulnerabilities, testing limits and the economics of food fraud.
1. Introduction: The Scope of Misrepresentation
Canadian inspectors identified fish labelled as Patagonian toothfish that was actually Antarctic toothfish, a lower-value species, according to the Canadian Food Inspection Agency’s (CFIA) food-fraud report for 2024–25 [1]. The report prices the two at about C$41.70 and C$23.17 per kilogram, using January 2026 figures from a price-tracking website. The difference is C$18.53 per kilogram at those approximate prices. By preventing the misrepresented fish from entering the marketplace, the CFIA says it protected consumers from paying the higher price for the lower-value species [1]. The gap is the easy part of the story. The harder question is where along the supply chain the label stopped being true, and a test cannot answer it by itself. The same report cautions that non-compliance may have arisen at various points in the supply chain, and that an unsatisfactory result does not always mean the problem began in the exporting country or where the product was sampled [1].
That distinction between detecting a problem and attributing it runs through this article. A laboratory result can establish that a sample failed the applicable test. It does not, by itself, identify where in the supply chain the problem arose or whether it was deliberate. The CFIA programme found 28 unsatisfactory results among 85 targeted samples of imported olive oil [1]. Because those samples were chosen for risk, the figure is not a market-wide fraud rate, but it shows that premium claims can fail when regulators test the products behind the labels. Each section that follows takes the gap into a different setting: what a sample can show, what analytical methods can and cannot resolve, how the question changes from species to quantity to contamination, and what drives the incentive.
Long supply chains can create more handovers at which a product’s composition, origin or declared quantity becomes harder to verify [2]. Processing can also remove visible clues, while regulators cannot test every product or shipment. Together, these conditions can make a detected discrepancy difficult to trace to its source.
The CFIA reported that its enforcement action in 2024–25 prevented more than 150,000 kilograms of misrepresented food from being sold in Canada [1]. The programme shows what testing and enforcement can detect, but its results cannot establish how much misrepresentation remains undetected across the wider market.
2. Regulatory Surveillance and Typology
Surveillance numbers depend on how samples were chosen. The CFIA uses two sampling categories. Marketplace monitoring, also called targeted surveys, is planned from risk intelligence and uses random sampling of prepackaged products at retailers by an independent third party under contract to the CFIA. Targeted inspectorate samples are taken by CFIA inspectors at importers, domestic processors and retailers, and they aim at higher-risk businesses, using factors such as a history of non-compliance, unusual trading patterns and gaps in preventive controls [1]. The CFIA states that targeted inspectorate results are not representative of overall compliance in the Canadian marketplace [1].
In 2024–25, the CFIA tested 886 samples for authenticity. The 246 marketplace samples covered two products, fresh meat tested for added sulfites and tea tested for dyes, and all 246 were satisfactory. Among the 640 targeted inspectorate samples, 82 (13%) were unsatisfactory and 8 (1%) were classed as investigative [1]. The two groups tested different products by different methods and cannot be compared directly, but the contrast shows how much a sampling design shapes what the numbers can say.
Among the targeted samples, 28 of 85 olive-oil samples, 15 of 81 honey samples and 6 of 37 grated-hard-cheese samples were unsatisfactory. All the olive oil was imported, and all 15 unsatisfactory honey results were among 68 imported samples [1]. As the introduction noted, though, the place where a sample is taken does not establish where the problem arose. The cheese findings concerned cellulose above permitted levels, a measure of composition rather than of intent.
The Codex Alimentarius Commission sets international food standards and is developing specific guidelines on preventing and controlling food fraud [3]. Those guidelines remain in draft development. The Commission adopted the text at Step 5 at its 47th session [3], and a 2026 circular letter (CL 2026/77-FICS), listed on the European Commission’s CCFICS page [4], circulated it for further comment at Step 6/7, according to the Codex circular-letter listing [19]. The responsible committee’s 28th session is scheduled for 12 to 17 October 2026 in Perth [4]. National authorities continue to apply their own laws and enforcement frameworks.
The table below is this article’s own organizing framework. Its examples are illustrative and are not findings from the sources cited elsewhere.
| Category | Definition | Illustrative Mechanism | Illustrative Impact |
|---|---|---|---|
| Species Substitution | Replacing a declared premium species with an inferior or lower-cost alternative. | Selling a lower-value fish as a higher-value one. | Consumer deception; possible safety implications, depending on the species and the consumer. |
| Dilution | Mixing a high-value liquid or bulk ingredient with a lower-cost liquid or filler. | Adding sugar syrups to honey, or cheaper oils to extra virgin olive oil. | Lower cost for the seller; altered chemical profile. |
| Origin Fraud | False declaration of a food’s geographic origin or provenance. | Misdeclaring a product’s origin to obtain preferential tariff treatment or exploit a reputation for quality. | Misleading origin claims; possible misuse of protected designations, including a Protected Designation of Origin (PDO). |
| Counterfeiting | Replicating the packaging, branding and trademarks of established products. | Copying brand labels, security seals and lot codes. | Direct brand damage; potential safety bypass. |
| Quantity Fraud | Deliberate misrepresentation of the net weight or volume of packaged goods. | Counting ice glaze as part of the declared weight of frozen seafood. | Inflated unit prices. |
| Production-Claim Fraud | Misrepresenting farming, harvesting or processing methods. | Labelling conventionally grown crops as organic, or caged eggs as free-range. | Erosion of trust in certification schemes. |
3. Honey: Industrial Syrups and Methodological Limits
Honey shows how hard it is to match a laboratory result to a legal finding. It also accounts for most of the quantity in Canada’s itemized enforcement list: 133,420 kilograms was honey that the CFIA classed as adulterated [1]. The CFIA’s itemized quantities are given in mixed units (kilograms, litres and an unlabelled count) and do not sum to its headline figure of more than 150,000 kilograms, so this article quotes the headline as the CFIA’s own and does not reconcile the two [1]. The European Commission’s coordinated action “From the Hives” examined imported honey for undeclared sugar syrups [5].
During the action, 147 of 320 sampled import consignments (46%) were flagged as suspicious for containing undeclared exogenous sugar syrups [5]. The pattern by country was uneven:
- China accounted for the highest absolute number of suspicious consignments [5].
- Türkiye also had a high proportion of its tested consignments flagged [5].
- All 10 consignments imported from the United Kingdom were flagged as suspicious. The Commission noted that traceability information suggested some honey may have originated elsewhere and been processed in the UK before re-export [5].
The Commission also stated a limit on what its testing could show: the methods identified chemical markers associated with exogenous sugars rather than quantifying how much syrup had been added [5]. The methods used to generate the reported results were outside the scope of ISO/IEC 17025 accreditation, a limitation relevant to legal certainty and enforcement [5].
A Government Chemist news item published on 27 March 2026 summarized a Grocer article describing disagreement over the reliability and interpretation of current analytical methods, including NMR. According to that summary, the Grocer article reported that variation in testing methods and the absence of harmonised approaches leave the extent of honey fraud unclear [2].
4. Olive Oil: Climate Shocks and Chemical Diagnostics
In the 2022/23 marketing year, the European Commission’s agriculture directorate forecast EU olive-oil production to fall by 39% to almost 1.4 million tonnes. It forecast Spanish production to fall by 56% to 663,000 tonnes, attributing the fall to the effect of drought on olive yields, and it said low production expectations had already pushed prices up [6]. Higher prices can raise the reward for adulteration, though the forecast says nothing about whether fraud increased.
Olive-oil fraud can involve adulteration with lower-value oils or misrepresentation of grade and production claims such as “extra virgin” [1]. The CFIA tests olive oil by gas chromatography to assess sterol and fatty-acid profiles [7]. That is a summary, not the whole analysis: the applicable method depends on the analytical question and the relevant standard.
The International Olive Council (IOC) maintains separate methods for different analytical questions [8]. The table lists the IOC methods behind three common measurements and what each can help assess.
| Parameter | IOC Method | What the Result Helps Assess | Target Analytes |
|---|---|---|---|
| Fatty Acid Composition | COI/T.20/Doc. No 33/Rev. 1 (2017): fatty acid methyl esters by gas chromatography [8] | Whether the fatty-acid profile is consistent with the declared oil. A foreign seed oil is one possible explanation for an inconsistent profile. | Palmitic, oleic, linoleic and linolenic acid proportions. |
| Sterol Composition | COI/T.20/Doc. No 26/Rev. 5 (2020): determination of the composition and content of sterols, triterpenic dialcohols and aliphatic alcohols by capillary column gas chromatography [8] | Whether the sterol profile is consistent with the declared oil. Interpretation depends on the thresholds in the applicable standard and on the suspected adulterant. | Campesterol, brassicasterol and beta-sitosterol, among others. |
| Ultraviolet Absorbance | COI/T.20/Doc. No 19/Rev. 5 (2019): spectrophotometric investigation in the ultraviolet [8] | Whether absorbance at 232 and 270 nm (K232, K270) is consistent with the applicable grade and quality criteria. Elevated values can be associated with oxidation or refining, but are not by themselves universal proof of adulteration. Interpretation depends on the applicable standards and the other analytical findings. | Specific extinction coefficients at 232 and 270 nm. |
A genuine declared blend, in which cheaper oils are mixed in and stated on the ingredient label, is different from fraudulent undeclared dilution, in which a lower-grade or adulterated product is sold at an extra virgin price.
5. Seafood: Species Substitution and Short-Weighting
Seafood shows how the verification question changes with what is being checked. Fraud takes two main forms here: species substitution in processed fillets, and short-weighting caused by excess ice glaze.
Species identity. The CFIA tests fish for species substitution by DNA barcoding, comparing the identified species with the species associated with the declared common name [7]. A result of that kind can say which species is in the package. It cannot say which handover made the label wrong; that is a question for the records behind the shipment. The toothfish price gap shows what is at stake for the buyer, which is the consumer’s loss from paying for the higher-value species and not the profit captured by any particular business. In fiscal years 2012 and 2013, the US Food and Drug Administration (FDA) ran DNA tests on wholesale seafood that it considered at the highest risk of mislabelling, including cod, haddock, catfish, basa, swai, snapper and grouper. It found the species correctly labelled 85% of the time [9].
Net weight. Species testing is separate from net-weight testing. On 2 September 2025, the FDA published results from a 2022–2024 sampling assignment covering 28 imported frozen seafood samples: 25 shrimp, two squid and one tilapia. The samples combined products targeted after complaints with general surveillance samples [10]. Ten samples (36%) were short-weighted by 2.3% to 9.9%, eight of them shrimp and two squid [10][11]. The FDA noted that glazing itself is common industry practice, but that short weighting can occur when the declared net weight includes the weight of the ice glaze, and that overstating net weight in this way is not permitted [10].
6. Spices and Powders: Chemical Adulteration and Public Health
Seafood fraud, as described above, concerns identity and quantity and mainly costs consumers money. Ground spices introduce a different problem, contamination, where chemical evidence can indicate a dangerous substance without showing who added it or why. Ground spices can also hide added pigments and contaminants that would be easier to see in whole products.
A 2024 study led by Stanford researchers collected 356 turmeric samples, 180 of dried roots and 176 of powder, from bazaars in 23 cities across India, Pakistan, Sri Lanka and Nepal. It found detectable lead above 2 μg/g in 51 of them [12]. Samples with lead at or above 18 μg/g had lead-to-chromium molar ratios near 1:1, which the authors described as suggestive of lead chromate, an industrial pigment [12]. In Patna, Karachi and Peshawar, lead exceeded 1,000 μg/g in some samples. The researchers modelled child blood lead levels from the measured concentrations and projected levels up to 10 times higher than the US CDC’s threshold of concern. Those blood-lead levels were modelled, not measured in people [12].
On 17 September 2025, the Food Safety and Standards Authority of India (FSSAI) wrote to its regional directors and to state and union territory food safety commissioners asking for an enforcement drive on whole and powdered spices from 1 to 31 October 2025. The drive was limited to manufacturing units and was meant to check compliance with prescribed parameters, including labelling. It followed surveillance between November 2024 and February 2025 that had found non-compliance in quality parameters, contaminants, microbial counts and labelling [13]. On 30 September 2025, FSSAI wrote that the 17 September letter was kept in abeyance until further orders and asked recipients not to initiate any action under it until further instructions were issued [14].
| Category | Non-Compliance Identified in Sources | Interpretive Note |
|---|---|---|
| Suspected deliberate adulteration | Lead and chromium ratios suggestive of lead chromate in some turmeric samples [12]. | The ratios point to a likely pigment source. They do not prove intent. |
| Quality, contaminant and labelling non-compliance | Quality parameters, contaminants, microbial counts and labelling, as identified in FSSAI surveillance before the directive [13]. | These findings should not all be treated as deliberate economic adulteration. |
7. Supply Chain Economics and Enforcement Mathematics
One possible explanation for food fraud is the financial incentive to substitute cheaper ingredients or misrepresent product quantity. Whether fraud is profitable in a particular case depends on the gains, the probability of detection, the penalties and the cost of losing inventory. Official seizure figures cannot supply most of these variables on their own.
A deliberately simple model shows how those pieces fit together. It treats one shipment, assumes that detected product earns no revenue and is seized, and ignores the profit the seller could have made by trading honestly. The model estimates a simplified expected return from the fraudulent shipment, before comparing it with the return from honest trading. It subtracts production cost and expected enforcement cost from expected revenue:
E(G) = P × V − C − (1 − P) × (F + S)
Here P is the probability that the fraud evades detection and V is the revenue if it does. C is the cost of producing the adulterated product, which is spent whether or not the fraud is detected and so sits outside the probability term. F is the fine if the fraud is detected. S is any further loss from seizure, such as disposal costs or lost contracts, beyond the production cost already counted in C.
All five variables are illustrative, because the article has no verified values for them. The model therefore identifies what drives the incentive and does not estimate actual returns.
To show how a small percentage scales in high-volume processing, consider a hypothetical plant handling 100,000 kilograms of ground beef a day with a 1% shortfall. For illustration, assume the shortfall is valued at C$16.00 per kilogram; this is a modelling input, not a verified market price:
- Daily shortfall volume = 100,000 kg × 0.01 = 1,000 kg
- Daily financial value = 1,000 kg × C$16.00/kg = C$16,000
- Cumulative value over 100 days = 100 × C$16,000 = C$1,600,000
Estimates that food fraud costs the global economy tens of billions of dollars circulate widely in media and industry discussions. No single, comprehensive measurement establishes the total value of food fraud across all commodities and jurisdictions. Fraud that is never detected or investigated may leave no confirmed enforcement record. Global estimates therefore depend on assumptions and often combine unlike categories, including safety violations, spoilage and trade disputes, with intentional deception. Defensible analysis relies on commodity-level testing and official enforcement records rather than unverified macro-economic totals.
Enforcement operations give a partial view. INTERPOL reported results it coordinated across 21 participating countries in Operation OPSON XI, which ran from December 2021 to May 2022: 1,100 tonnes of illicit food and drink seized, an estimated value of US$3,718,495, and 43 arrests [15]. Europol, which coordinated the EU-wide actions in the same operation, reported that it received results from 26 countries, including almost 27,000 tonnes of fake food seized and 80 arrest warrants [16][18]. The European Anti-Fraud Office (OLAF), whose targeted action was part of the same operation, reported about 14.8 million litres of illicit or counterfeit alcoholic beverages, wine and beer [17]. The figures cover different reporting bodies and country groups, so they are not directly comparable and are not combined here. Neither estimates the size of the illicit market or how much of it escaped detection.
8. Conclusion: The Verification-First Supply Chain Framework
Testing a finished product detects a problem only after the goods have been made, shipped and sometimes sold. Sampling and checking records earlier in the chain can reveal discrepancies before the product reaches the final buyer. Each control below needs a named actor, a point in the chain where it applies, and a trigger for follow-up.
| Control | Who Acts, and Where | What Triggers Follow-Up | What It Can and Cannot Show |
|---|---|---|---|
| Independent analytical testing | Regulators or accredited laboratories sample at import and at processor, blender and packer handovers, choosing targets by risk factors such as past non-compliance or unusual trading patterns [1]. | A failed result triggers follow-up to establish the nature and likely source of the non-compliance. Depending on the finding and applicable requirements, authorities may detain, remove, destroy or require relabelling of affected product [1]. | Detects chemically altered or mislabelled product. Methods such as DNA barcoding and chromatography are established. NMR may provide supporting evidence where the method has been validated for the specific authenticity question, and its interpretation is disputed in honey testing [2]. |
| Mass-balance accounting | Processors and blenders reconcile raw-input mass against finished-product mass for each lot, using expected yields and documented process losses such as moisture loss, trimming and by-products. Inspectors and auditors review the reconciliations. | A material discrepancy that remains unexplained after reconciliation prompts an investigation of that lot. | Flags material gaps that remain once legitimate process losses are accounted for. It does not say what caused the gap. |
| Auditable supplier records | Buyers and auditors check supplier records against invoices, shipping documents and customs data when each lot arrives. | A mismatch in origin, quantity or consignee puts the lot on hold until the supplier explains it. | Flags inconsistencies in origin and supplier records. Records alone cannot verify that a physical product is authentic. |
| Enforced commercial liability | Regulators and courts impose penalties, and regulators publish outcomes of enforcement action so that results are visible to other firms [1]. | A confirmed violation triggers enforcement under the applicable legal framework. Published outcomes can make enforcement more visible to other firms [1]. | Can reduce the reward from fraud when detection and penalties are credible. |
The controls address different failure modes, but their effectiveness remains unproven. The CFIA report documents follow-up on non-compliant results and six letters to foreign authorities concerning 32 non-compliance issues originating in their jurisdictions [1]. It does not report how often those follow-ups located the source. That share, alongside failure rates from comparable sampling programmes tracked over time, is the evidence that would show whether the controls work.
This article has not estimated detection rates or expected gains, so it proposes no specific penalty level. The model in Section 7 implies only that penalties and detection probability both matter to the incentive. Sampling, mass-balance checks and supplier records can narrow the search, and credible enforcement gives firms a further reason to prevent misrepresentation before it reaches the market. A failed test is where that investigation starts, not an account of how the failure happened.
Sources
[1] Canadian Food Inspection Agency, Food Fraud Annual Report 2024 to 2025, published 25 June 2026 and last modified 20 July 2026: https://inspection.canada.ca/en/about-cfia/science-and-research-cfia/our-research-and-publications/food-fraud-report-2025. Supports the toothfish example and its prices (taken by the CFIA from a price-tracking website, January 2026), the caution that non-compliance may have arisen at various points in the supply chain, the 886-sample total and the 246 and 640 split, the marketplace-monitoring products and results, the 640-sample outcome (82 unsatisfactory, 8 investigative), the olive-oil, honey and grated-cheese results and their level-of-trade breakdown, the representativeness limit, the CFIA’s headline of more than 150,000 kg prevented from sale and its itemized quantities, including 133,420 kg of adulterated honey (the itemized quantities are in mixed units and do not sum to the headline), follow-up on non-compliant results, the six letters on 32 non-compliance issues, and the published enforcement outcomes.
[2] UK Government Chemist, “Understanding honey fraud and the role of authenticity testing”, 27 March 2026: https://www.gov.uk/government/news/understanding-honey-fraud-and-the-role-of-authenticity-testing. Summarizes a Grocer article noting that complex global supply chains can limit traceability and increase uncertainty, that there is disagreement over the reliability and interpretation of analytical methods, including NMR, and that method variation and the lack of harmonised approaches leave the extent of honey fraud unclear.
[3] Codex Alimentarius (FAO/WHO), “CAC47 / Food fraud work advanced”: https://www.fao.org/fao-who-codexalimentarius/news-and-events/news-details/en/c/1727516/. Reports that the draft guidelines on the prevention and control of food fraud were adopted at Step 5 by the 47th Codex Alimentarius Commission, within the Codex eight-step procedure. The guidelines are not a final standard.
[4] European Commission, Codex Alimentarius page on the Codex Committee on Food Import and Export Inspection and Certification Systems (CCFICS), listing the 28th session (Perth, 12 to 17 October 2026) and circular letter CL 2026/77-FICS, “Draft guidelines on the prevention and control of food fraud” (the page lists the session and the circular letter; the Step 6/7 status is cited to [19]): https://food.ec.europa.eu/horizontal-topics/international-affairs/international-standards/codex-alimentarius/ccfics_en
[5] European Commission, “From the Hives” EU coordinated action on honey (2021–2022): https://food.ec.europa.eu/food-safety/eu-agri-food-fraud-network/eu-coordinated-actions/honey-2021-2022_en
[6] European Commission, DG AGRI, minutes of the Civil Dialogue Group on agricultural markets, 11 July 2023, item on olive oil and table olives: https://agriculture.ec.europa.eu/system/files/2023-07/cdg-agricultural-markets-2023-07-11-minutes_en.pdf
[7] Canadian Food Inspection Agency, Food Fraud Annual Report 2023 to 2024: https://inspection.canada.ca/en/food-labels/food-fraud/cfia-s-role/food-fraud-report-2024. Describes the olive-oil (gas chromatography) and fish (DNA barcoding) methods.
[8] International Olive Council, Standards, methods and guides (COI/T.20 methods of analysis, including Doc. No 19/Rev. 5 (2019), No 26/Rev. 5 (2020) and No 33/Rev. 1 (2017), with titles as printed in the methods): https://www.internationaloliveoil.org/what-we-do/chemistry-standardisation-unit/standards-and-methods/
[9] US Food and Drug Administration, “FDA DNA Testing at Wholesale Level to Evaluate Proper Labeling of Seafood Species”: https://www.fda.gov/food/guidanceregulation/guidancedocumentsregulatoryinformation/seafood/ucm419982.htm
[10] US Food and Drug Administration, “Sample Collection and Analysis of Imported Frozen Seafood for Economically Motivated Adulteration: Year 2022-24”: https://www.fda.gov/food/economically-motivated-adulteration-food-fraud/sample-collection-and-analysis-imported-frozen-seafood-economically-motivated-adulteration-year-2022
[11] US Food and Drug Administration, “FDA Releases Results on Economically Motivated Adulteration (Short Weighting) in Seafood”, 2 September 2025: https://www.fda.gov/food/hfp-constituent-updates/fda-releases-results-economically-motivated-adulteration-short-weighting-seafood
[12] Forsyth et al., “Evidence of turmeric adulteration with lead chromate across South Asia”, Science of the Total Environment, 2024, article 175003: https://doi.org/10.1016/j.scitotenv.2024.175003. Abstract also at Stanford FSI: https://fsi.stanford.edu/node/36768
[13] Food Safety and Standards Authority of India, letter RCD-02001/216/2025-Regulatory-FSSAI, dated 17 September 2025, “Enforcement Drive on Spices (Whole & Powder) for the month of October, 2025”. Published on the FSSAI website (fssai.gov.in, advisories section).
[14] Food Safety and Standards Authority of India, letter RCD-02001/216/2025-Regulatory-FSSAI, dated 30 September 2025 as printed on the letter (the file name refers to 1 October), keeping the 17 September letter in abeyance until further orders and asking recipients not to initiate action under it. Published on the FSSAI website (fssai.gov.in, advisories section).
[15] INTERPOL, “Food crime operations” (Operation Opson XI, INTERPOL-coordinated results from 21 participating countries): https://www.interpol.int/en/Crimes/Illicit-goods/Food-crime-operations
[16] Europol, “Food fraud: about 27 000 tonnes off the shelves”, 17 November 2022 (Operation OPSON XI results reported to Europol by 26 countries; the page did not return readable text in the verification check, so the figures rest on this listing and the secondary corroboration in [18]): https://www.europol.europa.eu/media-press/newsroom/news/food-fraud-about-27-000-tonnes-shelves
[17] European Anti-Fraud Office (OLAF), press release No 13/2022, 17 November 2022, on its targeted action within Operation OPSON XI: https://anti-fraud.ec.europa.eu/media-corner/news/148-million-litres-alcoholic-drinks-seized-across-europe-2022-11-17_en
[18] Food Safety News, “Huge amounts of dodgy food and drinks seized in global operation” (secondary report attributing the 26 countries, about 26,800 tonnes and 80 arrest warrants to Europol, used to corroborate [16]): https://www.foodsafetynews.com/huge-amounts-of-dodgy-food-and-drinks-seized-in-global-operation/
[19] Codex Alimentarius, circular letters listing, including CL 2026/77-FICS (request for comments on the draft guidelines on the prevention and control of food fraud at Step 6/7): https://www.fao.org/fao-who-codexalimentarius/resources/circular-letters/en/?y=2026
