Brussels bans what Beijing, London and the US food authority allow – the BPA can, even for export

Since 20 July 2026 no new drinks cans whose inner coating was made with bisphenol A may be placed on the market in the European Union. What was delivered before that may still be filled until 20 July 2027.^[1][2] After that no drinks can with a BPA inner coating may be filled in the EU.

One might take this for a detail of food chemistry. I take it for an object lesson in how regulation comes about in Europe: from the strictest value available, without regard to the fact that other expert authorities arrive at an entirely different result, and with consequences far beyond the internal market.

A factor of 1,000, among experts

In 2023 the European Food Safety Authority, EFSA, determined that a person can take in 0.2 nanograms of BPA per kilogram of body weight a day without a health risk being expected. Previously the value stood at 4 micrograms, that is 20,000 times higher.^[4] The German Federal Institute for Risk Assessment considers 200 nanograms appropriate and contradicts the derivation of the EFSA value on several points.^[5] The British expert committee COT has adopted the German value.^[6] The EU Commission based its ban on EFSA’s assessment that BPA could harm the immune system.^[3]

The disagreement is no trifle. The BfR considers even the endpoint unsuitable on which EFSA bases its value: a change in particular immune cells in mice. No health hazard for humans can be predicted from that, it says.^[5] The COT takes the same view and considers effects in humans up to the German value unlikely.^[6] The US food authority FDA writes that BPA is safe in the quantities that occur in food today; coatings containing BPA in metal cans remain permitted there.^[7]

To be clear: the factor of 1,000 does not mean that 1,000 times more BPA may be in a can in the United Kingdom. It describes how differently expert authorities assess the same body of studies. That is precisely why it is politically so explosive.

The risk is not invented

Anyone declaring BPA harmless is making it too easy for themselves. The Endocrine Society, the world’s largest professional association in hormone research, expressly supported the EU ban.^[11] A study published in 2026 found a reduction in insulin sensitivity of around 9 per cent in 40 young adults after five days of targeted BPA administration.^[12] The dose, however, was 50 micrograms per kilogram of body weight, 250 times the German and 250,000 times the European value. The study did not examine long-term illness or normal consumption from cans.^[12]

Neither “all perfectly safe” nor “every can makes you ill” is therefore an honest summary. What is honest is this: science is in dispute about the relevant dose. And where science is in dispute, politics has to give reasons for choosing the strictest reading and say what it costs.

Outside Europe hardly anyone is following

Those who have followed are the neighbours who align their food law with the EU anyway: Norway and Iceland via the EEA, Switzerland by its own decision.^[15][16] Outside Europe no state on this list has a comparable ban. Stricter is only a single American state, Washington, and even that only for drinks cans.^[17]

European bottlers can still export, to every country in the world. But only in the can that Brussels prescribes. The regulation prohibits not only sale but already the use of BPA in manufacture, and every can a filler buys in the EU is placed on the market here. There is no exemption for export goods.^[1][2] Anyone shipping drinks from Germany to America or Asia therefore bears the cost of the switch even where local competitors may carry on using the authorised BPA can.

This concern is not mine alone. In a British consultation on a ban of its own, participants expressly warned of competitive disadvantages and trade barriers, because the United Kingdom and the EU would thereby diverge from the United States, China and Japan.^[10]

Where two European expert authorities are a factor of 1,000 apart, that is no proof for a ban. It is an obligation to give reasons.

Who earns from the switch

Every switch has winners. AkzoNobel has put 32 million euros into a new plant in Vilafranca in Spain that makes bisphenol-free coatings for the can industry.^[13] The European coatings association CEPE has a BPA working group of its own; in March 2025 its board included representatives of PPG, Sherwin-Williams and AkzoNobel.^[14]

The association itself writes that its talks with the authorities led to a “workable outcome” for can and heavy-duty coatings.^[14] The regulation does indeed contain exactly one permanent exemption for coatings: for containers of more than 1,000 litres.^[1] But the same report also shows that CEPE contradicted EFSA’s assessment and fought for adequate transitional periods, not for a ban.^[14] That the coatings lobby brought the ban about is not supported by the documents. That lobbying took place, and who did it, is.

What I expect

Protecting health needs robust science, and regulation needs proportionality. Anyone forcing a whole supply chain to switch has to disclose three things: what additional protection the ban brings, what it costs companies and consumers, and why the differing assessment of two expert authorities was rejected. We found no such calculation for the BPA can.

For me this is an example of regulation that burdens Europe’s drinks industry without anyone doing the sums. More on European packaging regulation is in my open letter to the federal chancellor and in the background piece Verpackungsverordnung aussetzen.

Andreas W. HerbFounder and CEO MBG Group, Paderborn

What I expressly do not claim

I do not claim that BPA is harmless, nor that export to countries outside the EU is impossible; drinks in BPA-free cans may be sold everywhere. How much the switch costs per can has not been published by anyone; that it burdens the sector is my assessment. I do not claim that the coatings industry bought the ban. The factor of 1,000 describes differing assessments by the authorities, not differing limits for a can.