European food retailers: a horsemeat scandal that exposed long meat supply chains

The situation

On 14 January 2013 the Food Safety Authority of Ireland (FSAI) told its counterparts in England that it had found horse DNA in beef products sold by Tesco, Iceland and Lidl. The next day it published the results. Of 27 beef burgers it had tested, 10 contained horse DNA and 23 contained pig DNA. Nine of the ten horse-positive samples held only traces. The tenth, a Tesco Everyday Value burger, held enough to suggest that horsemeat made up about 29% of its meat relative to beef. The burgers came from three plants: Liffey Meats and Silvercrest Foods in Ireland, and Dalepak Hambleton in Yorkshire. They had been sold through Tesco, Dunnes Stores, Lidl, Aldi and Iceland.

The FSAI’s chief executive, Professor Alan Reilly, said the products posed no food safety risk. He also said there was “no clear explanation at this time for the presence of horse DNA”. That second remark set the problem for everyone downstream. A retailer can recall a batch of burgers in a day. Explaining how meat from a horse got into a product labelled beef, and whether it happened by accident or by design, takes much longer and reaches into suppliers the retailer may never have met.

Over the following four weeks the story moved from cheap burgers to frozen ready meals, from Ireland to France, Luxembourg, Cyprus, the Netherlands and Romania, and from food inspectors to the police. Retailers, brand owners and governments had to decide how far to widen their recalls, whom to keep buying from, how much testing to demand and who would pay for it. This case follows those decisions and asks what they reveal about long processed-meat supply chains in which the name on the pack belongs to a company several steps away from the abattoir.

Background

Who sold the products

The products at the centre of the affair were cheap processed beef lines, mostly sold under retailers’ own labels or under a frozen-food brand. Tesco’s Everyday Value range and Aldi’s Today’s Special range were value ranges designed to sit at the lowest price point in their categories. Findus, a frozen-food brand, sold beef lasagne in the UK that it did not make itself. Tesco was the largest grocer in Britain: in the four weeks to 17 February 2013 Kantar Worldpanel put its share of the grocery market at 29.7%.

None of these companies slaughtered cattle for these products. They bought finished goods from manufacturers, who bought frozen meat from processors and traders, who in turn bought from abattoirs. Retailers set specifications and audited their direct suppliers. The specification said beef. Whether the meat was beef depended on paperwork passed along the chain.

The manufacturers and processors

Silvercrest Foods, in Ballybay, County Monaghan, was part of ABP Food Group, one of Ireland’s largest meat companies. It made burgers for retailers and for Burger King. Comigel was a French maker of frozen ready meals. Its subsidiary Tavola ran the factory at Capellen in Luxembourg that produced beef lasagne and spaghetti bolognese for several European brands and retailers, including Findus and Tesco. Comigel bought its meat from Spanghero, a French meat processor that belonged to the Poujol group.

Behind Spanghero sat traders. According to an investigation by the Organized Crime and Corruption Reporting Project (OCCRP), the Romanian abattoir Doly Com sold more than 350 tonnes of horsemeat in 2012 to Draap Trading, a company registered in Limassol, Cyprus, and owned through a British Virgin Islands company. The price was about €2 a kilo. The meat was stored in a cold warehouse in the Netherlands before it went on to French buyers, including Spanghero. Jan Fasen, a Dutch trader whom the OCCRP said had introduced Doly Com’s owner to the Poujol group in October 2012, had been convicted in the Netherlands in January 2012 of selling South American and Canadian horsemeat as Dutch and German beef. He was appealing and denied wrongdoing.

Why horsemeat

Horsemeat is legal food in much of continental Europe, and it is cheaper than beef. A Reuters report carried by CNBC in February 2013 said Romanian abattoirs paid about 3.5 lei a kilo for a horse and about 5.5 lei a kilo for a cow, at a time when many small Romanian farmers still relied on horses for fieldwork. The Romanian exporter CarmOlimp said horsemeat made up less than 1% of its turnover. Romanian officials said their own checks cleared the two implicated abattoirs of relabelling horsemeat as beef, and a French inquiry later concluded that the meat had left Romania correctly labelled as horse. The gap between the price of horse and the price of beef, multiplied across hundreds of tonnes, was the incentive. The long chain of traders was the opportunity.

The industry

A concentrated retail market buying from a fragmented supply base

British grocery retail was dominated by four chains. In February 2013 Kantar recorded overall grocery sales growing 3.7% a year, behind grocery price inflation of 4.3%, which meant volumes were falling. Discounters were gaining ground: Aldi reached a record 3.3% share and Waitrose, at the other end of the market, a record 4.8%. Kantar’s director Edward Garner read this as a sign that the market was polarising between premium and discount shoppers. In this setting the big chains competed hard on price in value ranges, and they had strong bargaining power over manufacturers of own-label goods.

A Porter-style reading of the supply side helps. Retail buyers were few and large. Manufacturers of frozen ready meals were more numerous and could be switched, so their margins were thin and they were under constant pressure on input costs. Meat, the largest input, was a commodity traded across borders by many small intermediaries, some of them little more than a company registration and a bank account. Information ran badly down that chain. Each buyer knew its seller but rarely the seller’s seller. The threat that mattered was not a new entrant or a substitute product. It was the ease with which a cheaper material could be passed off as the specified one, because routine checks were designed to catch unsafe food rather than mislabelled food.

Regulation and enforcement

Food law in the EU required food to be safe and correctly labelled, and required operators to trace products one step back and one step forward. Enforcement was national. In England, a change in 2010 split food policy. Responsibility for food authenticity and labelling where there was no safety issue moved from the Food Standards Agency (FSA) to the Department for Environment, Food and Rural Affairs (Defra) and the Department of Health. Local authorities kept the job of enforcement. The FSA remained responsible for safety.

The National Audit Office (NAO) later set out how stretched that system had become. England had 490,308 registered food businesses in 2012. Government spent about £241 million a year protecting consumers from food incidents, three-quarters of it through local authorities. Local authority food sampling had fallen by 26% since 2009-10. The number of public analysts, the qualified scientists who test food for enforcement, had fallen from 40 in 2010 to 29, and four of the official control laboratories had closed in two years. Some 63% of local authorities had cut food law enforcement staff since 2008-09. Meanwhile reports to the national food fraud database had risen by about two-thirds since 2009, to 1,380 in 2012. UK authorities had last tested for horsemeat in beef products in 2003. In 2012, one in six products tested for the presence of a different species, other than horse, had failed.

Technology

DNA testing for species was well established and not expensive per sample. The FSAI decided in November 2012 to test beef products, using a method it had developed to check the composition of meat products. The test detects the presence of horse DNA and can estimate its proportion. It cannot say on its own whether low levels came from cross-contamination in a shared plant or from deliberate substitution. That limit shaped the arguments that followed, because most positives were traces.

Exhibit 1: Food enforcement capacity in England before the scandal

Source: National Audit Office, Food safety and authenticity in the processed meat supply chain, October 2013.

The decision

The first results raised no safety alarm, but they did raise questions of honesty and control. The people who had to act were spread across several organisations, each with its own options.

Retailers and brand owners

Retailers could withdraw only the batches in which horse had been found, or clear every product from the implicated plants as a precaution. A narrow withdrawal protected sales and avoided implying that other lines were suspect. A wide one cost more in the short term, but it limited the risk of a second, worse discovery. Tesco moved quickly and publicly. Its group technical director, Tim Smith, said the company would not tolerate any compromise in the quality of its food. When the problem later reached ready meals, Tesco withdrew its Everyday Value spaghetti bolognese as a precaution before its own test results came back, because the meal came from the same factory as Findus lasagne.

The second choice was what to do about suppliers. Retailers could stand by a long-standing supplier while it investigated, or drop it. Dropping a supplier sent a clear message to shoppers but could leave shelves empty and might punish a company that had itself been deceived. Burger King, which had more than 500 outlets in the UK and Ireland, chose to drop Silvercrest within days.

The third, slower choice was structural. Retailers could keep buying through long chains of traders at the lowest price, or shorten their chains and buy closer to home, which would cost more. Tesco’s chief executive, Philip Clarke, later chose the second path in public, while admitting he could not promise it would not raise prices.

Governments and regulators

The UK government had to decide whether to treat the burgers as an Irish problem, to be handled by the FSAI and the companies concerned, or to demand testing across the whole market. Wider testing would show the size of the problem but would take resources from local authorities that had already cut sampling. A second question was whether this was a matter for food inspectors or for the police. On 11 February, the Environment Secretary, Owen Paterson, said no case for criminal action had yet been found in the UK, while the FSA said it was working closely with the police in case that changed. Labour’s shadow environment secretary, Mary Creagh, wrote to him suggesting that the contamination might be the result of systematic adulteration.

The European Commission faced a similar choice at EU level: leave testing to member states, or co-ordinate a common programme and bring in Europol. Within the industry, there was also a question of who paid for testing. MPs on the Environment, Food and Rural Affairs (EFRA) Committee later argued that retailers, not consumers, should bear that cost.

What they did

Phase one: the burgers (mid to late January 2013)

Retailers named by the FSAI removed the implicated batches. Silvercrest withdrew all its products and suspended production while ABP investigated. Millions of burgers were taken off sale over the following month. On 24 January Burger King dropped Silvercrest as a supplier for its UK and Irish restaurants. Irish officials traced the horse DNA at Silvercrest to a third-party supplier of beef in Poland. The Irish agriculture minister, Simon Coveney, said the positive burgers correlated directly with that supplier. ABP’s chief executive, Paul Finnerty, said the company had never knowingly bought or traded in horse products and apologised. ABP promised new plant management, DNA testing and independent audit of third-party suppliers. Liffey Meats and Dalepak were cleared by further tests.

Phase two: the ready meals (early February 2013)

The second wave was far more serious. On 4 February Findus UK withdrew its beef lasagne after a French supplier alerted Comigel. On 7 February the FSA confirmed that some Findus beef lasagne contained up to 100% horsemeat. Tesco had already withdrawn its frozen Everyday Value spaghetti bolognese, also made by Comigel. Its tests found that most positive samples held traces below 1%, but three showed horse DNA above 60%. A test for phenylbutazone, a veterinary drug not permitted in the food chain, was clear. Tim Smith said the levels suggested Comigel had not followed the proper production process and that Tesco would not take food from its facility again. He added: “We are very sorry that we have let customers down.”

Aldi recalled its Today’s Special frozen beef lasagne and spaghetti bolognese in the UK after its own tests found horsemeat making up between 30% and 100% of the meat in some lasagne samples. The FSA said the Aldi and Findus results suggested “gross negligence or possibly criminality”. In France, Switzerland, the Netherlands and Sweden, retailers pulled ready meals, and within days horsemeat had been found in products in Germany, Norway and Austria. On 18 February Nestlé withdrew two beef pasta products in Italy and Spain and one in France after finding more than 1% horse DNA.

Phase three: tracing the chain and calling in the police (mid-February 2013)

Comigel stopped buying from Spanghero on 8 February. Spanghero said it had bought the meat as beef from a Romanian supplier and threatened to sue. Romania’s prime minister, Victor Ponta, denied that any Romanian company had broken European rules. On 14 February the French government said Spanghero had knowingly sold about 750 tonnes of horsemeat labelled as beef over six months, and its licence was suspended. Spanghero said it had acted in good faith. On 18 February, after talks with its unions, the French government allowed it to resume making minced meat, sausages and ready meals, but not to hold frozen meat. The same week, British police arrested three men at meat plants, and the UK found phenylbutazone in six horse carcasses.

The European Commission called in Europol on 13 February and urged member states to carry out DNA tests on beef products and checks for phenylbutazone in horsemeat. It agreed a co-ordinated control plan, run from 11 February to 4 April at an estimated cost of €2.5 million. In the UK, the FSA and Defra required industry to test all processed beef products and send the results to the FSA. The programme covered products supplied to schools, hospitals, prisons and the armed forces. By 15 February, 29 of the first 2,501 UK beef products tested had more than 1% horsemeat.

Phase four: reviewing the system (February 2013 to September 2014)

On 27 February Philip Clarke told the National Farmers’ Union conference that Tesco would bring meat production closer to home. Tesco’s pledges included sourcing all fresh chicken from UK farms from July 2013, offering two-year contracts to farmers who wanted them, DNA testing all processed beef, publishing test results online and setting up an independent panel to hold it to account. Clarke said: “I hope it doesn’t mean prices increases but I can’t stand here and tell you that it won’t.”

Parliament and government then looked at the system. The EFRA Committee reported in July 2013. The NAO reported on 10 October 2013. Defra and the Department of Health asked Professor Chris Elliott of Queen’s University Belfast to review the integrity of food supply networks. He published an interim report in December 2013 and a final report, published on 4 September 2014.

What happened

Shoppers

The effect on sales was immediate and concentrated in the categories involved. In the four weeks to 17 February 2013, Kantar Worldpanel found UK frozen burger sales down 43% and frozen ready meal sales down 13% on a year earlier. Tesco’s market share slipped 0.4 points to 29.7%, but Kantar’s Edward Garner warned against blaming it all on horsemeat, because Tesco had not repeated a large promotion from the year before. Iceland, a frozen-food specialist named in the first Irish results, still grew sales by 10.1%.

Exhibit 2: UK grocery market in the four weeks to 17 February 2013

Source: Kantar Worldpanel, as reported by Retail Gazette, February 2013.

Testing

The wider testing showed a serious problem in a small number of supply chains, not a market-wide one. Across the EU, national authorities carried out 4,144 DNA tests on beef products in the 2013 programme and found more than 1% horse DNA in 193 of them, or 4.66%. They ran 3,115 tests for phenylbutazone on horsemeat and found it in 16, or 0.51%. Food businesses reported a further 7,951 DNA tests, with 110 positives (1.38%). France reported the most non-compliant samples, 47 of 353. A second EU round in 2014, extended to Norway, Iceland and Switzerland, found horse DNA in 16 of 2,622 samples, or 0.61%. The Commission did not publish the level of horse in each positive sample, so the results do not separate deliberate substitution from cross-contamination.

In the UK, by the end of August 2013 industry had tested 24,480 samples and found 47 positive for horse DNA, across 17 product lines. Twenty-eight local authorities tested 514 beef products and found two with horse DNA and four with pig DNA. The NAO reported that all 150 UK samples submitted to the EU programme were negative. One trade report of the EU figures listed 14 UK non-compliant samples, which does not match the NAO’s account. This case follows the NAO, which had direct access to the UK data.

Exhibit 3: Horse DNA testing results, 2013 and 2014

Sources: FSAI via Shelflife (January 2013); European Commission test results page; NAO (October 2013). UK rates calculated from the NAO’s counts.

Reports and reforms

The EFRA Committee concluded that the contamination was fraud, carried out through what its chair, Anne McIntosh, called “a complex network of companies trading in and mislabelling beef”. It said retailers had not been vigilant enough where meat was traded many times, and recommended that they carry out regular DNA tests and report the results to the FSA, at their own expense. The NAO recommended that government reconsider the split of responsibilities between the FSA, Defra and the Department of Health, improve the sharing of intelligence and watch laboratory capacity.

The Elliott Review set out eight pillars of food integrity: consumers first, zero tolerance, intelligence gathering, laboratory services, audit, government support, leadership and crisis management. Its main structural proposal was a food crime unit. The National Food Crime Unit (NFCU) was set up inside the FSA in December 2014.

Courts

The criminal cases took years and ended mostly in other countries. In the Netherlands, a court in Den Bosch sentenced the meat trader Willy Selten in April 2015 to two and a half years in prison, the first jail term of the affair, for selling about 300 tonnes of horsemeat as beef. Of 167 samples taken from his stock in February 2013, 35 tested positive for horse DNA. Selten said he had been careless, not a deliberate fraudster. In England, Southwark Crown Court sentenced the slaughterhouse owner Peter Boddy and his manager David Moss in 2015 for traceability and invoice offences. Boddy was fined and Moss received a suspended sentence. The judge said they were sentenced only for what they had admitted, not for substituting horse for beef, although there was strong suspicion of wider complicity.

In France, the trial of two former Spanghero managers and two Dutch traders ended in April 2019. Jacques Poujol, Spanghero’s former director, was sentenced to two years, 18 months of them suspended, banned from the meat trade for two years, and ordered to forfeit €100,000 found at his home. The Dutch trader Johannes Fasen received two years in prison. Patrice Monguillon, the Spanghero plant manager, and Hendricus Windmeijer, who worked with Fasen, received one-year suspended sentences.

Exhibit 4: Main legal outcomes

Sources: Food Safety News (March 2015, April 2015, April 2019).

Other views

Retailers as victims or as part of the cause

Retailers presented themselves as victims of fraud by suppliers they had trusted, and in a narrow legal sense the courts agreed: no retailer or brand owner was prosecuted. Their critics argued that the retailers’ buying model made fraud more likely. The EFRA Committee said retailers had not been vigilant enough. The National Farmers’ Union, whose members stood to gain from shorter chains, said consumers wanted more British meat. Its president, Peter Kendall, called for shorter supply chains sourcing from British farmers, though he accepted that imports could work if retailers were diligent about how they bought. A year after Tesco’s pledges, a Farmers Weekly investigation found that Tesco had stopped working with 30 beef processors and abattoirs and had given two-year contracts to about 200 lamb and 100 beef farmers. It also found that the share of British beef on Tesco’s shelves had fallen to 82% in November 2013, and that some suppliers complained of imposed contract changes. Tesco replied that the pledge was about shortening links in the supply chain, not about percentages.

Suppliers and governments in the chain

ABP said it had been let down by a third-party supplier and never knowingly bought horsemeat. Spanghero said it had acted in good faith. Romania’s government rejected any suggestion that its abattoirs had relabelled meat, and Ponta said he was very angry. The French investigation supported Romania on that point. These accounts are consistent with the evidence that the fraud happened in the trading and processing links in western Europe rather than at the point of slaughter.

Regulators and the state

The NAO’s findings were in effect a criticism of the state’s own preparedness: fragmented responsibility, falling sampling and a decline in laboratory capacity. The EFRA Committee also said the FSA’s responsibilities had been unclear. After the convictions in France, the campaign group Foodwatch said the sentences were not severe enough to deter others, pointing out that Spanghero’s horsemeat trading had brought in €1.9 million of turnover between February and December 2012. The OCCRP estimated Spanghero’s earnings from the scheme at more than €500,000 over six months. Those figures sit beside a two-year sentence and a €100,000 confiscation.

Academics

A 2017 paper in npj Science of Food by Stephanie Brooks, Chris Elliott and colleagues found that testing and surveillance had become routine in industry, and that companies were sharing intelligence among themselves through the Food Industry Intelligence Network. But they were still not sharing it with regulators, mainly for reasons of confidentiality and competition. The authors also found that the NFCU’s annual budget was about £900,000, against the £2 million to £4 million a year the Elliott Review had recommended.

Reading it through a framework

Porter’s value chain, extended to the supply network

Michael Porter’s value chain divides a firm’s activities into primary activities (inbound logistics, operations, outbound logistics, marketing and sales, service) and support activities, one of which is procurement. In the horsemeat case the failure sat in procurement and inbound logistics, but not inside the retailer’s own chain. It sat in the chains of companies two, three and four steps upstream.

Laying the Findus and Tesco ready meals out as a series of linked value chains makes the point. Romanian abattoirs sold horse at about €2 a kilo. Draap Trading in Cyprus resold it, a Dutch cold store held it, Spanghero in France processed it, Comigel’s Tavola factory in Luxembourg cooked it into lasagne and bolognese, and Findus, Tesco and Aldi sold it. Each link captured a margin. None of the routine checks tested species. Retailers had invested heavily in downstream activities (brand, marketing, store operations), where they earned their returns. Upstream they relied on specifications and documents.

The response can be read as an attempt to pull control back into the procurement activity. Tesco dropped Comigel’s facility and 30 beef processors and abattoirs, introduced DNA testing of all processed beef and offered longer contracts to farmers. In value chain terms, Tesco spent more on procurement to buy assurance that the product matched the label. The EFRA Committee’s recommendation that retailers pay for testing is the same argument made by a regulator: the firm whose brand captures the value should bear the cost of protecting it.

Stakeholder mapping by power and interest

A power and interest grid explains how fast the response escalated. Before January 2013, most stakeholders had low interest in species authenticity. Shoppers assumed beef was beef. Regulators had moved authenticity away from the FSA and cut sampling. Retailers audited suppliers mainly for safety.

The Irish results moved shoppers sharply up the interest axis. Each shopper had little power alone, but together they cut frozen burger sales by 43% in a month, which gave them influence that retailers could not ignore. Retailers had high power over suppliers and high interest in their brands, so they recalled stock and dropped suppliers within days. Governments had high formal power but, as the NAO showed, limited capacity. Their interest jumped once ministers faced questions in Parliament, and they used their power to require industry testing rather than doing the testing themselves. Police and prosecutors entered once the FSA spoke of possible criminality. Suppliers such as Spanghero and Comigel lost power almost entirely once their customers and governments turned against them.

The grid also shows who was missing. Consumers of food in schools, hospitals and prisons were covered by the testing programme only because the FSA asked public bodies to check their own contracts. Farmers, through the NFU, used the moment to press for shorter chains, a position that served their commercial interests as well as their stated concern for integrity.

Since then

Horsemeat fraud did not stop in 2013. In July 2017 Europol announced the results of an investigation in which Spanish police arrested 65 people and Belgian police arrested the suspected leader, a Dutch businessman linked to the 2013 Irish burger case. The group was accused of slaughtering horses that were old, sick or marked unfit for consumption, forging their identification documents and microchips, and selling the meat mainly outside Spain. The case showed that horse identification and passport systems remained weak.

The NFCU grew slowly. For years it lacked the power to apply for search warrants and had to rely on the police and local authorities. From 1 May 2025 its officers in England and Wales gained the ability to apply for and carry out search warrants under the Police and Criminal Evidence Act. Its head, Andrew Quinn, said the powers would let investigations proceed more directly. Chris Elliott said the ability to search premises straight after arrests would make it harder for criminals to dispose of evidence.

Foodwatch’s warning in 2019 that similar incidents would recur cited the contamination of eggs with the insecticide fipronil in 2017, a different problem that also crossed several European borders. The split of responsibilities that the NAO criticised in England, and the question of who pays for authenticity testing, remain matters of policy debate rather than settled arrangements.

What to take from it

• A retailer’s brand carries the risk of every company in its supply chain, including traders and processors it has never dealt with directly.
• Checks designed only for safety leave room for fraud, so authenticity needs its own tests, its own budget and a clear owner in government.
• Fast, wide precautionary withdrawals cost sales in the short run but protect the brand against a second and worse discovery.
• The price gap between a cheap material and the specified one, multiplied across hundreds of tonnes, is a predictable incentive that buyers should price into their assurance spending.
• Cross-border food fraud is slow to prosecute, and sentences that are small relative to the gains do little to deter it.

Questions to think about

1. How much responsibility should a supermarket carry for fraud committed by a supplier three or four steps away from it, and how should that be reflected in law?
2. Spanghero was said to have sold about 750 tonnes of horsemeat as beef. If horse cost about €2 a kilo and beef cost twice that, what was the approximate gross gain, and how does it compare with the penalties handed down in 2019?
3. UK industry tested 24,480 samples and found 47 positives. Was this testing programme good value, and how would you decide how much testing is enough after the crisis has passed?
4. Tesco pledged to buy more British meat, yet a year later the share of British beef on its shelves had fallen. How would you judge whether its response was substantive or mainly reputational?
5. Should authenticity policy sit with the Food Standards Agency, with a farming ministry, or with a crime agency? Set out the strengths of each.
6. Most positives in the EU programme were low-level. How should a regulator communicate results that cannot distinguish cross-contamination from deliberate fraud?

Concepts at work

Value chain analysis, stakeholder mapping, supply chain risk. Explore them in Insights.

Further reading

Oxford University Press covers this story as a teaching case in Cases in International Business on Oxford Business Trove. The sources below are the public record this page was written from.

Written from the public sources listed below. Figures are as those sources report them.