A 2023 Australian parliamentary inquiry asked for a national online gambling regulator and a phased total advertising ban — as of 2026-08-20 neither exists, and the narrower bill written instead has passed only one chamber

What is happening?

In 2023 a standing committee of the Australian House of Representatives reported on online gambling harm in a report titled You Win Some, You Lose More, carrying 31 recommendations. The two heaviest were a phased total ban on online gambling advertising across all media and the creation of an independent national online gambling regulator. As of 2026-08-20 neither of those two exists.

The formal government response arrived on 2026-05-12, about three years after the report. It did not take up the national regulator. In place of a total advertising ban it set out restrictions by time of day and by context.

The package those restrictions came from had been announced earlier, on 2026-04-02. Television advertising would be limited to three spots an hour between 06:00 and 20:30, prohibited outright during live sport broadcasts inside that window, and radio advertising would be prohibited during school travel times. Online advertising would reach only logged-in adult account holders, who could opt out. Advertising using celebrities or athletes would be prohibited, as would advertising on stadiums and playing uniforms. Commencement was announced for 2027-01-01, before any bill existed.

The legislation followed on 2026-07-02: the Interactive Gambling Amendment (Gambling Reform) Bill 2026, alongside the National Self-exclusion Register (Cost Recovery Levy) Amendment Bill 2026. The first inserts a new Part 7C into the Interactive Gambling Act 2001, converting what had been co-regulatory industry codes into statutory prohibitions, with civil penalties up to AUD 364,000 for an individual, AUD 1,820,000 for a body corporate and AUD 2,730,000 for breach of the anti-avoidance provision. The enforcement agency named is the Australian Communications and Media Authority.

The bill went to the Senate Environment and Communications Legislation Committee, with submissions closing 2026-07-24 and a reporting date of 2026-08-17. After that stage 16 amendments were agreed between the major parties, covering inducement marketing, a fourteen-day bar on inducement advertising to new accounts, a three-month bar on inducement advertising to a person who has left the national self-exclusion register, a prohibition on performance-linked commissions for staff of wagering operators, a new opt-out advertising register named Adstop, and an obligation to identify customers showing signs of gambling problems and stop marketing to them.

The amended bill passed the House of Representatives on 2026-08-18, with two members recorded as voting against. As of 2026-08-20 the Senate has not voted on it.

Whose problem is this?

The body that recommended an independent regulator has no power to create one, and the body that could create one declined to. Nothing in that arrangement produces a second attempt or a date by which the question must be answered again.

Where does this problem end?

The boundary here separates an instrument from an institution. The advertising rules are moving, slowly and in reduced form, through a process that can be dated at every step. The regulator is not moving at all, because after the response of 2026-05-12 no process is running that would produce one.

What is the state now, and what should it be?

Now

Should be. The target state in this case was written by a parliamentary committee and it is unusually specific. The report of 2023 asked for a total prohibition on online gambling advertising across all media, phased in over three years, and an independent national regulator with the standing to oversee online gambling as a whole. That is the standard on the record, produced by a committee of the parliament rather than by an outside body. The government then set a second and narrower target of its own, an effective date of 2027-01-01 for the advertising restrictions it did adopt. Both targets can be checked. What no source opened here supplies is a date, a condition or a review point at which the regulator question would be asked again, which means the first of the two targets currently has nothing attached to it at all.

How big is it?

What can be stated is the denominator and two disputed magnitudes. The national population was 27,801,023 at 2025-12-31, published by the national statistical agency on 2026-06-18. That is the base against which any participation or harm rate would be applied, and no rate that could be applied to it was confirmed here.

One opened source, a statement by a national body representing medical practitioners, puts annual gambling losses at about AUD 31.5 billion and at more than AUD 1,500 per adult, and describes that as the highest per-person loss in the world. That page carries no publication date that could be confirmed here; it describes two years having passed since the 2023 report, which places it at about 2025. No opened source reconciles that total with any other, and no opened source states whether it covers all gambling or online gambling alone.

One opened commentary states that one million people either experience serious gambling harm directly or live with someone who does. It names no underlying statistical source for the figure, so it is carried here as a claim made by that source rather than as a measured quantity.

The distance between those two kinds of number is the reason this block is marked. A loss total measures money leaving households in a year. A harm count measures people. Neither of the figures above was traced to a primary series in this round, and multiplying a confirmed population by an unconfirmed rate would produce a number that looks measured and is not.

Needs a new measurementno source opened here gives a count of the people affected that can be traced to a primary prevalence series, and the loss totals that circulate do not agree with one another. The population file for this document records `not-derivable` rather than a manufactured interval.

Under what conditions does it arise?

1. The rules sat in industry codes rather than in statute. Under a co-regulatory scheme the content of the rule and the point at which it is enforced are separated, and the enforcement agency acts on the code rather than on a direct statutory prohibition. The bill introduced on 2026-07-02 is what would move those prohibitions into the Act itself, with penalties attached, and until it passes both chambers the earlier arrangement is what remains in force.

2. Authority over gambling is split by level of government. Online wagering and its advertising sit with the Commonwealth, while poker machines and venue-based gambling sit with the states and territories, and parts of the 2026-05-12 response were referred back to those jurisdictions. A recommendation for a single national regulator is a recommendation to build the body that this split does not currently produce, which is also why no existing agency can be asked to deliver it.

3. A recommendation carries no clock. Nothing obliged the government to respond to the 2023 report by any particular date, and about three years passed before it did. Nothing obliges a second response on the recommendations that were not adopted, and no source opened here names a review point.

4. The revenue side is concentrated and countable and the harm side is neither. Advertising inventory has a price, a buyer and a contract term. The harm the restrictions are meant to reduce is spread across households, arrives later than the exposure, and as this document records has no agreed national count at all.

5. A bill that has passed one chamber is not law. The commencement date of 2027-01-01 was announced in April 2026, before introduction, and the instrument that would give it effect had cleared one of two chambers as of 2026-08-20. Announced dates and enacted dates are different objects, and only the second binds anyone.

What has been tried?

Two things have been tried here and they are not the same thing. One is an advertising instrument, which has been drafted, amended, widened by sixteen agreed changes and carried through one chamber, and which can be tracked by date at every step. The other is an institution, which was recommended once and then left out of the response, and which has had no process attached to it since.

What was found?

Why is it still unsolved?

Institutional gap — the advertising instrument is being built and the institution that was recommended alongside it is not being built, and nothing in the process requires anyone to revisit the second.

The first part of the gap is that a recommendation is not a deadline. The committee reported in 2023 and the formal response came on 2026-05-12, and no rule fixed anything in between. When the response arrived it adopted part of what had been recommended and left the regulator out, and at that moment the recommendation stopped having a live process attached to it. There is no next step in the sequence. A committee report does not lapse and it does not persist either. It stays on the record in exactly the same condition whether it is being acted on or not, which means that from the outside a recommendation under consideration and a recommendation abandoned look identical.

The second part is that the split between levels of government is what a national regulator would have been for. Online wagering advertising is federal, venue-based gambling is state and territory, and parts of the response were referred back to the states and territories. A body that sits across that line is precisely the thing that cannot emerge from either side of it on its own, since neither level has jurisdiction over the whole subject. So the recommendation that most needed a decision at the centre is also the one with the fewest actors able to take it forward without that decision.

The third part is that the instrument that is moving absorbs the attention that the institution would need. The advertising bill has dates, amendments, a committee stage and a division in the chamber, so it produces a continuous record of activity. The regulator has none of those and generates nothing to report. When one half of a recommendation set becomes a bill and the other half becomes silence, progress on the first is legible and the absence of the second is not, and the two are easily read as one movement.

The fourth part is that the reduced form can be described accurately as an advance. The bill does convert co-regulatory codes into statutory prohibitions with penalties attached, and the sixteen agreed amendments widened it further. That is a real change against the position before it. It is also not what was recommended, and both statements are true at once, which is the condition under which a gap becomes stable rather than merely open.

What observation would mean it is solved?

Candidates — (a) the bill passes the Senate, commences, and the enforcement agency publishes what it did with the new powers (b) an independent national body with statutory functions over online gambling exists and reports publicly (c) measured gambling harm in the population falls on a series that is published on a fixed schedule.

(a) is the nearest and it answers only half the question. Passage would settle the advertising instrument and would say nothing about the regulator, which is the other recommendation and the one this document is named for. Commencement is also not enforcement. The measure of whether Part 7C works is what the enforcement agency does with the three penalty tiers after 2027-01-01, and a statute that is never applied looks the same in the statute book as one that is.

(b) is the observation that would show the gap itself had closed, because it is the thing that has never existed here. It is also the least likely of the three to appear, since the response of 2026-05-12 declined it and no process is currently running that would produce a second decision. If it did appear, the test would be whether the body has functions over the whole subject or only over the federal part of it, since a national regulator confined to one side of the split would leave the condition that produced the recommendation unchanged.

(c) is the outcome measure and it cannot currently be read at all. As this document records, there is no agreed national count of people experiencing gambling harm, the loss totals in circulation disagree, and the one prevalence study named in the sources is a pilot. A measure that has no baseline cannot register a fall. Establishing the series would itself be a precondition for using this test, and the three have to be read together, with (b) as the one that would say whether anything structural had changed.

What is it connected to?

Fills with researchstate and territory regulation of poker machines and venue-based gambling, the relationship between wagering sponsorship and the financing of professional sport, credit and payment controls on online wagering accounts, the operation and coverage of the national self-exclusion register, and the same recommendation-to-implementation gap in other Australian parliamentary inquiry subjects. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• Which of the 31 recommendations were accepted, which were accepted in principle, which were noted and which were declined. This is the central missing item of the document. The response document itself is the published position of the department that issued it, and it did not open here, so every statement above about what the response did rests on media summaries of it rather than on the classification.
• Why the interval between the report and the response was about three years. No source opened here carries an on-the-record explanation from any minister, department or committee. Sources record the interval and its endpoints, and none of them supplies a reason.
• On what stated grounds the national regulator was not adopted. The sources record the outcome. None of them records an argument attributed to a named public body against creating one, or a description of what was considered in its place.
• The exact publication date of the 2023 report. The parliamentary page that carries it refused automated requests. The year and the count of 31 recommendations were confirmed from two other opened sources, and the day and month were not.
• A reconciled annual loss total. One opened source gives about AUD 31.5 billion with no stated basis year and no statement of whether it covers all gambling or online gambling alone. No opened source reconciles it against any other figure, and none states its counting basis.
• The primary source for the figure of one million people. The commentary that states it does not name one, and no statistical publication opened here carries the same quantity.
• What the Senate committee reported by its due date of 2026-08-17. The committee report was not read here, so what it recommended and whether the sixteen agreed amendments followed from it are both unestablished.
• What enforcement of the new Part 7C would cost or require. No opened source states what resources the enforcement agency would receive, how many matters it expects, or how compliance with the time-of-day limits would be monitored.
• The national prevalence baseline. The one prevalence study identified in this round is described as a pilot and its page refused automated requests, so no figure from it appears here.

See the evidence

No Australian government document was read directly except one statistical release. The population figure comes from the statistical agency page, which opened, and it is the only high-graded quantitative row here. Everything about the 2023 report, the 2026-05-12 response and the current status of the bill reaches this document through secondary sources, and the four blank rows above are the primary documents that would have settled them: the committee report, the government response, the prevalence pilot and the parliamentary bill tracker all refused automated requests or timed out, and each is left blank with the reason recorded rather than filled with a substitute link. One of those blanks holds the central missing item, which is how each of the 31 recommendations was classified in the response. Where sources overlap they agree: the committee referral dates appear identically in two opened sources, the House passage date of 2026-08-18 and the two votes against appear in two, and the absence of a national regulator from the response appears in three. Where a figure is weak it is marked and quarantined: the annual loss total of about AUD 31.5 billion and the figure of one million people are graded low, are attributed in the text to the source that states them, carry no basis year in either case, and support no claim in the state table, the population file or the headline. The population file records not-derivable for that reason rather than combining a confirmed population with an unconfirmed rate. This is a Path A output, so observation_refs is empty and provenance_mode: press-derived.

People affected

Estimated range Not derivable

The reason and what is missing are listed under “What is missing” below