Australia's Two Opposite Food Label Decisions of 2026

In the space of seven weeks in early 2026, Australian and New Zealand food regulators made two decisions that point in opposite directions.

On 13 February 2026, food ministers agreed that the Health Star Rating — the front-of-pack scoring system that has been voluntary since 2014 — should become mandatory. The reason was blunt: the industry had been given a target of displaying the rating on 70 percent of eligible products by November 2025, and had reached 39 percent in Australia and 36 percent in New Zealand.

On 31 March 2026, Food Standards Australia New Zealand abandoned a long-running proposal to require added sugars to be declared separately on the nutrition information panel. Its assessment concluded that the change offered no clear public health benefit, might mislead or confuse consumers, and would impose substantial costs on industry and government.

One decision says voluntary labelling failed and must be compelled. The other says a specific mandatory disclosure is not worth compelling. Read together they are a useful case study in how food labelling policy is actually made — and a reminder that "more information on the packet" is not automatically better.

The Health Star Rating assigns a score from half a star to five stars, in half-star increments, displayed on the front of the pack. Understanding the algorithm explains both its usefulness and its most criticised results.

The score is calculated per 100 grams or 100 millilitres, and the calculation is category-adjusted — dairy products, fats and oils, and beverages are scored against different scales rather than all competing on one. That adjustment exists because comparing olive oil to breakfast cereal on a single scale produces nonsense.

Two design consequences follow, and both have driven the criticism the system attracts.

The rating is comparative within a category, not absolute. A four-star cereal is a better cereal, not a health food. Consumers frequently read the star count as an absolute endorsement, which the system was never designed to provide.

And because fibre and protein add stars, products can improve their rating through fortification rather than through reduction of sugar, salt or fat. This is the same structural incentive that operates in the United Kingdom's nutrient profiling model, and it attracts the same objection: reformulating to game a score is not the same as reformulating to improve a food.

The Health Star Rating was introduced in 2014 as a voluntary scheme, on the theory that competitive pressure would drive adoption. Manufacturers of products scoring well would display their ratings, consumers would come to expect the label, and products without a rating would look conspicuous by its absence.

That theory did not survive contact with the market, and the reason is straightforward once stated.

A voluntary front-of-pack rating is adopted selectively. A manufacturer with a four-star product displays it. The same manufacturer's 1.5-star product in the next aisle carries no rating at all. Because the label is optional, its absence carries no penalty and communicates nothing definite to a shopper — who cannot distinguish "this product scored badly" from "this company does not participate".

The result is a system that appears informative while being structurally biased. Every rating displayed is, on average, better than the ratings not displayed. Adoption stalling at 39 percent in Australia and 36 percent in New Zealand — against a 70 percent target — is precisely the outcome that self-selection predicts.

Federal Assistant Health Minister Rebecca White framed the decision in exactly these terms: the industry had failed to meet the voluntary target it had been set. With the target missed by a wide margin over more than a decade, the ministers' options were to extend the voluntary period again or to mandate. They chose to mandate.

This is the single most transferable lesson in this article for any country considering front-of-pack labelling. Voluntary schemes produce selective disclosure, and selective disclosure is worse than no scheme at all in one specific respect: it creates an appearance of transparency that the underlying data does not support.

The second decision is more surprising and requires careful handling, because it cuts against the direction most food labelling policy has taken.

Proposal P1058 would have required added sugars to be declared as a separate line in the nutrition information panel, distinct from total sugars. The distinction matters nutritionally: the sugar naturally present in milk or whole fruit is not treated the same way in dietary guidance as sugar added during manufacturing. The United States has required this declaration since 2016.

FSANZ abandoned the proposal on 31 March 2026. Its stated reasoning had several strands. The assessment concluded that including added sugars in the panel offered no clear public health benefit for consumers or for their ability to make choices consistent with dietary guidelines. It raised the possibility that the information might mislead or confuse consumers, reduce trust in food labelling, or prompt purchases inconsistent with dietary guidelines. And it found that mandating the declaration would impose substantial costs on industry and on government.

The confusion argument is the one that deserves scrutiny, because it is counterintuitive. The concern is that a product with a low added-sugar figure but high total sugars — a fruit juice, for instance — could be read as low in sugar when it is not. Meanwhile a product with moderate added sugar might be judged more harshly than its overall profile warrants. Whether consumers actually make these errors in practice is an empirical question on which evidence exists in both directions.

It is worth being precise about what was not abandoned. Amendments under a separate proposal, P1062, concerning voluntary nutrition content claims about added sugars, are unaffected. Manufacturers may still make added-sugar claims; they simply will not be required to declare the figure.

Whether this was the right call is genuinely contested. Public health bodies have criticised it. FSANZ's assessment is public and its reasoning can be read directly rather than through summaries.

No front-of-pack system has escaped criticism, but the Health Star Rating has attracted a specific and recurring complaint: certain products score in ways that strike shoppers as obviously wrong. Understanding why is the best way to understand what the rating can and cannot tell you.

The recurring pattern involves sweetened products that score respectably because of offsetting components. A flavoured milk drink earns points for protein and calcium that partially offset its sugar. A sweetened breakfast cereal earns points for fibre. A fruit-based snack earns points for fruit content even when that fruit has been concentrated into something closer to confectionery. In each case the algorithm is working exactly as designed, and the result still surprises people.

The mirror-image complaint concerns whole foods that score modestly. Olive oil, cheese, nuts and avocado all contain fat, and energy and saturated fat reduce stars. Category adjustment mitigates this — oils are scored against other oils — but it does not eliminate the counterintuitive results entirely.

There is also the participation problem discussed above, which produces a particular absurdity while the scheme remains voluntary: the products most likely to score badly are the products least likely to display a score, so the worst performers are invisible within the system meant to identify them.

None of this means the rating is useless. Within a category, comparing two cereals or two yoghurts, it generally directs shoppers toward the better option, and that is the use case it was built for. The failure is one of interpretation: a summary score invites absolute reading, and the system cannot support absolute reading.

This is why the algorithm review question matters so much for the mandate. Making a disputed score compulsory, without revisiting the offsetting rules that generate the disputed results, would make every one of these anomalies more consequential rather than less.

Australia's mandate decision places it in a growing group. The systems differ more than most coverage acknowledges.

The systems split into two philosophies, and the split matters.

Summary scores — Health Star Rating, Nutri-Score — compress everything into one symbol. They are easy to compare at a glance but hide the reasons behind the score, and they allow one good attribute to offset a bad one.

Nutrient-specific warnings — Chile's octagons, UK traffic lights — flag individual problems. They cannot be gamed by offsetting, and the warning models in particular have the strongest evidence base: Chile's mandatory warnings, combined with marketing restrictions, produced measurable reductions in purchases of flagged products.

Australia has chosen to mandate a summary score. That is a defensible choice given a decade of existing consumer familiarity, but it means adopting the format with the weaker evidence base and the known offsetting problem.

Australia's decision is not isolated. Across several jurisdictions the same sequence has repeated, and it is predictable enough to be worth stating as a pattern.

Stage one is a voluntary scheme introduced with industry cooperation, usually as an alternative to regulation. Industry supports it precisely because it is voluntary, and governments accept it because it avoids a legislative fight.

Stage two is partial adoption. Products that score well display the label. Products that score badly do not. Adoption plateaus somewhere well below full coverage — 39 percent in Australia's case, against a 70 percent target.

Stage three is a monitoring report that documents the shortfall, usually against a target the industry itself accepted.

Stage four is a mandate, justified by the documented failure of stage two.

The United Kingdom's traffic light system has lived in stage two for years, with high but incomplete voluntary adoption. Nutri-Score in Europe is fragmented, adopted by some member states and resisted in others, with disputes over whether it treats traditional regional foods fairly. Chile skipped the voluntary stage entirely and went straight to mandatory warnings — and has the strongest evidence of any system for changing purchasing behaviour.

The United States is currently at a variation of stage one: the FDA issued a proposed front-of-package rule in early 2025 focused on added sugar and sodium, and its 2026 agenda commits to summarising public comments and preparing options for a final regulation. Public comment reportedly showed broad support for front-of-package labelling, with disagreement concentrated on content and design rather than on the principle.

The inference for anyone watching this space is that voluntary front-of-pack labelling has now failed often enough, in enough countries, that regulators increasingly treat the voluntary stage as a delay rather than a genuine policy option.

The February decision was a direction, not an implementation. The process from here has several stages and will take time.

FSANZ has been asked to prepare a proposal to mandate the Health Star Rating within the Australia New Zealand Food Standards Code. Two rounds of public consultation are planned before a final regulatory decision, and FSANZ is leading engagement on how a mandatory system should operate in practice — including transitional arrangements for industry and consumer education campaigns. A decision is expected in 2027, with implementation and transition periods after that.

Several questions are genuinely open and will be settled through that process. Whether the underlying algorithm is revised at the same time as mandating it is the most consequential — mandating a disputed algorithm entrenches its known weaknesses. How long manufacturers have to change packaging determines the real start date. Which product categories are exempt determines coverage. And what enforcement looks like determines whether compliance is real.

Separately, FSANZ has consulted on introducing a single annual commencement date for its own labelling changes, proposing 30 May each year. This is administratively dull and practically significant: packaging changes are expensive, and allowing manufacturers to batch them into one annual transition lowers the cost of every future labelling reform.

Because Australia has now declined to separate these, it is worth understanding exactly what the distinction is and how to work around its absence.

Total sugars, the figure declared on nutrition panels in most countries, includes every mono- and disaccharide present, regardless of origin. That means the lactose in milk, the fructose in whole fruit, and the sucrose a manufacturer stirred in are all counted in the same number.

Added sugars covers only what was introduced during processing — table sugar, syrups, honey, concentrated fruit juices used as sweeteners. Dietary guidelines in most countries set limits on added or free sugars specifically, not total sugars, because sugar consumed within an intact whole food behaves differently: it arrives with fibre, water and a food matrix that slows absorption.

The practical workaround where no added-sugar line exists is comparison within a category. Find the plain version of the product and note its total sugars, then compare the flavoured version. The difference approximates what was added. Plain yoghurt against fruit yoghurt, unsweetened against sweetened plant milk, plain oats against flavoured instant oats — in each case the gap is informative.

This technique is imperfect, because flavoured products may also contain fruit that contributes sugar. But it is considerably better than reading total sugars alone and concluding that milk and cola belong in the same category.

Three transferable conclusions, regardless of where you live.

Treat any voluntary front-of-pack rating as self-selected. If a scheme is optional in your country, the products displaying it are systematically better than the products not displaying it. The absence of a rating is information, and it is usually unfavourable information.

Read summary scores as comparative, not absolute. A star rating, a letter grade or a colour code tells you how a product compares to others in its category. It does not tell you the product is good for you in any absolute sense. Four-star crisps are still crisps.

Check total sugars, not just added sugars. This is the practical consequence of the Australian decision. In markets without an added-sugar line, the total sugars figure per 100g remains available and comparable. And in markets with an added-sugar line, the total figure still matters — a product low in added sugar can be high in total sugar.

The broader pattern across every jurisdiction examined here is consistent. Regulators have concluded that voluntary schemes underperform, and they are moving toward mandates. Australia's February decision is one instance of a global trend, and the seven-week gap before the added-sugar reversal is a useful reminder that the trend is not uniform — each individual disclosure requirement is fought on its own evidence.

There is a final point worth making about how to read decisions like the added-sugar abandonment. It is tempting to interpret a regulator declining to mandate a disclosure as capture by industry, and sometimes that interpretation is correct. But it is not the only available reading, and treating it as automatic makes it impossible to evaluate any individual decision on its merits.

The alternative reading is that labelling requirements have real costs, that consumer comprehension of additional panel lines is genuinely uncertain, and that adding information is not costless if the information is misread. FSANZ published its reasoning, and the reasoning is testable — the claim that separate added-sugar figures could mislead consumers is an empirical claim that research can support or refute.

The useful posture for a reader is neither automatic trust nor automatic suspicion. It is to note what the regulator claimed, note that the claim is checkable, and treat the star rating mandate and the added-sugar abandonment as two separate decisions that happen to have been made by the same system weeks apart — because that is what they are.