The sugar levy, and the ethics of taxing what people drink

Dentistry panels like this topic because it sits exactly where their subject meets public policy. Tooth decay is the most common reason a young child in England is admitted to hospital for a general anaesthetic, sugar frequency is its main driver, and here is a national policy aimed squarely at one source of it — with several years of evidence behind it. That is a rare combination, and it is why a candidate who can talk about the levy in terms of mechanism rather than headline stands out.

The mistake is to treat it as a debate about taxation. The levy is better understood as a reformulation policy that happens to use the tax system as its lever. Once you see it that way, the ethics get sharper too, because the person whose choice was most constrained was the manufacturer, not the drinker.

The design detail that matters is the two thresholds. A drink with less than 5g of added sugar per 100ml pays nothing. From 5g it pays the lower rate, and from 8g the higher one. The rates were set per litre and have been uprated since, but the exact pence matter far less than what the steps did to behaviour.

A manufacturer facing a threshold has three choices: pay, pass the cost on, or reformulate to land just under the line. Most chose the third, and many did so between the announcement in 2016 and the start date in 2018 — before a penny had been collected. Tracking published by Public Health England found that the sugar content of drinks in scope fell by roughly a third per 100ml over that period, while the sugar in categories outside the levy barely moved. The levy raised revenue, but its main effect was on the product, not the price.

Two exemptions shaped the argument that followed. Milk-based drinks were excluded on the grounds of their calcium and protein, and pure fruit juices because their sugar is not added. Both exemptions have been contested since, and a 2025 government consultation proposed bringing milk-based drinks into scope and lowering the first threshold. Check the current position before an interview: the design is being actively revised.

The claim that matters to a dentist is not that the levy raised money. It is that fewer children needed teeth removed under general anaesthetic afterwards, and that the evidence for that is reasonable rather than perfect.

Researchers at Cambridge compared hospital admissions for tooth extraction due to decay before and after the levy, using the years before as the counterfactual. Their analysis, published in 2023, estimated a relative reduction of around a tenth in such admissions among children up to 18, with the largest fall in the youngest age groups — the children whose exposure is most likely to come from drinks rather than from anything they buy themselves. A parallel line of work found reductions in obesity among older primary-school girls, with no comparable effect in boys, which is the kind of uneven result that tells you the mechanism is real but not the whole story.

Say what that evidence is and is not. It is an interrupted time-series, not a trial: there is no group of English children who lived through 2018 without the levy. The estimates are associations with plausible mechanisms behind them. Presenting them as certainties overclaims; presenting them as nothing dismisses the best evaluation of a sugar policy anywhere. The honest line is that the effect is measurable, modest, and concentrated where the harm is worst.

This is where the marks are, and the trick is to run the pillars against the levy as it actually is rather than against a caricature of a sugar tax.

Autonomy. The objection is that the state is steering what adults drink. The reply is that the levy left every product on the shelf and constrained the manufacturer’s recipe more than the consumer’s choice; the drinker’s autonomy was barely touched, and the child’s — who has none in this transaction — was protected. Beneficence. The benefit is real, measured and concentrated in the youngest. Non-maleficence. The main harm alleged is economic and regressive: a flat levy takes a larger share of a poorer household’s income. That is worth conceding, then answering: reformulation meant most households paid little more, and the health harm being prevented also falls hardest on poorer households. Justice. A population measure reaches the children whose parents do not read labels, which targeted education never will.

Notice that the strongest objection is not about liberty at all. It is about whether a policy that reaches everyone is fair to the people who were never part of the problem. Candidates who can name that, and then say why they still support or oppose the measure, are arguing at the level a dental school wants.

This arrives in three shapes. The direct one: "Should we tax sugary drinks?" The disguised one, where the levy is never mentioned but the answer needs it — "How would you reduce childhood tooth decay?" or "Is it right for the government to interfere in what people eat?" And the scenario one: a parent objects that the state has no business in their shopping basket.

For the direct question, give the mechanism before the opinion. For the disguised question, use the levy as your worked example of a population measure with evidence. For the scenario, concede the autonomy point and then move the frame to the child.

Start with the government’s own page on the levy for the thresholds and current rates, then read the Cambridge evaluation of hospital extractions rather than the coverage of it — the methods section tells you exactly what kind of evidence it is. If you want the argument in the profession’s words, the British Dental Association has campaigned on sugar for years.

Two pieces here carry the surrounding context. Children’s tooth decay and hospital extractions is the problem the levy was aimed at, and water fluoridation is the other population measure you will be asked to compare it with. For the ethics framework itself, our guide to the four pillars is the place to rehearse.