Renata Cho works three shifts a week on the service desk at a suburban Melbourne supermarket, on top of a part-time payroll job, and her combined income sits at about $42,000 a year. Under the tax cut Parliament had already legislated for 2024 — a change she had heard about vaguely as "the big one coming" — she was scheduled to receive precisely nothing. The original design only touched marginal rates above $45,000; below that line, her tax bill was never going to move, no matter how the numbers above her changed. Then, from July 2024, an extra few hundred dollars a year started landing in her pay — real money, in a year when her rent and her grocery bill had both risen sharply. She got that money only because the government tore up the plan she was never going to benefit from and wrote a different one.
That she benefited is not in dispute. How the government got there is. For the better part of two years, Anthony Albanese and Treasurer Jim Chalmers told voters, journalists and Parliament, without qualification, that Labor's position on the Stage 3 tax cuts had not changed and would not change. Then, in January 2024, it changed completely. Every bracket was redrawn. The question this topic puts to the record is not only whether the new brackets were better policy — on the evidence, for most taxpayers, they clearly were — but whether a promise this explicit, repeated this often, can be broken this cleanly without a lasting cost to what any political promise is worth.
What Labor promised, then redesigned
The 2019 design: one big bracket, tilted to the top
The Stage 3 tax cuts were not Labor's creation. They were the third and final stage of a three-part personal income tax plan first legislated by the Morrison Coalition government in 2018 (Treasury Laws Amendment (Personal Income Tax Plan) Act 2018), then locked in and expanded after the Coalition's re-election, in the Treasury Laws Amendment (Tax Relief So Working Australians Keep More Of Their Money) Act 2019, passed with Labor's support in July 2019. Stage 1 and Stage 2 took effect earlier and mainly adjusted the lower and middle brackets. Stage 3, scheduled to start on 1 July 2024, was the biggest structural change of the three: it abolished the 37% tax bracket outright and merged the 32.5% and 37% bands into a single 30% rate applied to every dollar of taxable income between $45,000 and $200,000.
The effect was to flatten the middle of the income tax scale. Under the rates that applied before Stage 3, a worker's marginal tax rate rose from 32.5% to 37% at $120,000, then to 45% at $180,000. Under the original Stage 3 design, that same worker would have paid a flat 30% on every dollar from $45,000 up to $200,000, before jumping straight to 45%. Because the design did not touch the tax-free threshold or the 19% bracket below $45,000, nobody earning under $45,000 stood to receive a single dollar from Stage 3 as legislated in 2019. The benefit was concentrated above $120,000 and reached its ceiling, in dollar terms, for every taxpayer earning $200,000 or more — each of whom was set to receive the same maximum cut, worth an estimated $9,075 a year (approx.; calculated from the 2019 Act's brackets against the rates that applied before Stage 3), whether they earned $200,000 or $2 million. That flat ceiling above $200,000, and the zero benefit below $45,000, was the single most contested feature of the design from the outset — criticised by economists, the Parliamentary Budget Office's longer-run distributional work and figures inside Labor itself, years before Labor ever won government.
By the standards of comparable countries, the original design was unusually blunt. Several OECD income tax systems use four, five or more brackets across a comparable span of middle and upper-middle incomes (approx.; comparative OECD tax data). Compressing $45,000 to $200,000 into a single 30% Australian rate meant a worker on $50,000 and a worker on $190,000 would have faced an identical marginal tax rate on their next dollar earned — a flattening of the tax scale with few direct parallels in Australia's own postwar tax history.
Why Labor promised not to touch it
The "no change" pledge did not appear from nowhere. Bill Shorten's Labor went into the 2019 election flagging concerns about the then-proposed Stage 3 cuts, arguing they were too heavily skewed toward high income earners. Labor lost that election — an outcome that post-election analysis, including a review process Labor itself commissioned, linked partly to a platform many voters judged too complicated and too focused on raising revenue rather than cutting tax (approx.). When the full three-stage package returned to the Senate weeks after the Coalition's re-election, in July 2019, Labor voted for it in full rather than risk being blamed for blocking a tax cut a second time.
Anthony Albanese, who became Labor leader after that defeat, carried the lesson forward into 2022: neutralise tax as a political attack line by promising, unconditionally, to keep what had already been legislated. It worked as electoral strategy — tax barely featured as an issue in the 2022 campaign that put Labor into government. That same strategic logic is what makes the January 2024 reversal land as more than an ordinary policy adjustment. The commitment was not incidental or accidental. It was built, over several years, to be as close to unbreakable as a pre-election promise can be, precisely so voters would stop worrying about it. Breaking a promise engineered that carefully to be trusted carries a different weight than walking back an offhand pledge made once during a campaign.
Almost two years of "no change"
Labor fought the 2022 election on a clear position: Stage 3 would proceed exactly as legislated. Anthony Albanese and Jim Chalmers repeated that position, in near-identical language, at regular intervals through 2022 and 2023, including in a series of television and radio interviews in which both men were asked directly whether the government would revisit the design — and both said no. As late as December 2023, senior ministers were still ruling out change in public, only weeks before the position reversed entirely (approx.; contemporaneous reporting, Dec-2023–Jan-2024). This was not one passing remark that could later be read as aspirational. It was a sustained, repeated, unqualified commitment, made by the two people with the authority to change it, over a period long enough that voters, employers and financial planners had every reason to treat it as settled law in waiting.
The persistence of that commitment is not only a matter of political memory; it is written into the government's own official documents. Both the May 2023 Budget and the December 2023 Mid-Year Economic and Fiscal Outlook — the two most recent formal statements of the government's fiscal plans before the reversal — built their forward estimates on the original, unaltered Stage 3 design (Budget Paper No. 1, 2023-24, May-2023; MYEFO 2023-24, Dec-2023, approx.). That is the ordinary and correct way to budget for a policy already locked into law, but it also means the government's own primary fiscal record shows it still planning on the original numbers only weeks before it tore them up.
The reason the commitment mattered so much is tied to the same flaw that made the original design so contested. Cost-of-living pressure built sharply across 2022 and 2023, as inflation and interest rates rose faster than wages could keep up (see this masthead's inflation and real-incomes analyses). A tax cut worth its maximum only above $200,000, and worth nothing below $45,000, looked increasingly out of step with a squeeze that was hurting middle and lower earners hardest. That tension — a legislated design skewed to the top, sitting alongside a cost-of-living emergency concentrated further down the income scale — is the pressure that eventually broke the promise.
25 January 2024: the redesign
On 25 January 2024, at the National Press Club in Canberra, Albanese announced that the government would rewrite Stage 3 before it took effect. The new design, legislated within weeks as the Treasury Laws Amendment (Cost of Living Tax Cuts) Act 2024 and in force from 1 July 2024, kept the same basic shape of a bottom rate, a middle rate, an upper-middle rate and a top rate, but moved every threshold and cut the bottom rate itself. The 19% rate that applied to income between $18,201 and $45,000 was cut to 16%. The new 30% rate was set to run from $45,001 to $135,000, not to $200,000 as originally legislated. The 37% bracket, which the 2019 design had abolished entirely, was kept alive, now applying between $135,001 and $190,000. And the top 45% rate was moved down to start at $190,001, rather than $200,001.
The bill passed both houses of Parliament within about three sitting weeks of the announcement. The Coalition, having initially attacked the reversal on trust grounds, ultimately voted for the redesigned brackets rather than be seen opposing a tax cut that benefited most of its own voters (approx.; parliamentary proceedings, Feb-2024). Treasury described the change as broadly cost-neutral against the original Stage 3 costing over the medium term — an estimated cost in the order of $300 billion across the decade to the early 2030s either way (approx.; Budget Paper No. 2, 2024-25, May-2024) — which meant the argument was never really about how much the government would spend on tax relief. It was about who received it.
Nothing changed about the arithmetic between 2019 and 2024. What changed was how much political room there was to fix it.
The strongest case for the policy
A fairer redesign, delivered without extra cost, when households needed it most
Put the government's case at its best, because parts of it are simply true. The original Stage 3 design was legislated in 2019, in an entirely different economic environment — before COVID-19, before the sharpest run of interest-rate rises in a generation, before inflation cut deeper into essentials than into discretionary spending. A tax schedule calibrated for one set of conditions is not obliged to survive unchanged into a materially different one, and no government is required to stay locked into a specific rate structure simply because a predecessor legislated it five years earlier and an opposition once promised not to touch it.
On its own terms, the redesign delivered exactly what its defenders say it did. Every one of the roughly 13.6 million Australian taxpayers who stood to receive a cut under the original design still received one under the redesign (approx.; Treasury, Jan-2024) — nobody's total tax bill rose relative to the law that applied before Stage 3 existed at all. And an estimated 11.5 million of them — the large majority, concentrated below the roughly $146,486 crossover point — received a larger cut than they had been promised, not a smaller one (approx.; Treasury, Jan-2024). At a moment when two years of inflation had eaten into real wages and rate rises were squeezing every mortgage holder in the country, directing more relief toward workers earning $40,000 to $135,000, and somewhat less toward the roughly 1.2 million taxpayers earning above $200,000 who were best placed to absorb a smaller gain, is a defensible act of redistribution, not a betrayal of one (approx.; Treasury, Jan-2024). Treasury's own costings put the redesigned package at broadly the same cost to the budget as the original — nobody paid more tax to fund somebody else's bigger cut. The money simply moved to where the pressure was worst.
The evidence: the brackets, the crossover and the ledger
None of this needs to be taken on trust. The bracket thresholds and rates are set out in black letter law, in the 2019 Act and the 2024 Act that replaced its Stage 3 schedule, and the tax cut at any given income can be calculated directly from them against the rates that applied before Stage 3 was legislated at all. Treasury also published its own distributional tables alongside the January 2024 announcement, breaking the changes down by income band, state and electorate.
Who gained, who gave something up
The clearest way to see the redesign's actual effect is to compare, at a range of incomes, what each design delivers relative to the brackets that applied before Stage 3 existed. Below $45,000, the original design gave nothing at all, while the redesign — because it cut the 16% rate on income between $18,201 and $45,000 — delivered a genuine, if modest, cut to every taxpayer above the tax-free threshold: about $654 a year (approx.) for someone on $40,000, rising to about $804 a year (approx.) at exactly $45,000. Through the middle of the income scale the redesign stays ahead: at $60,000 it delivers roughly $1,179 against the original design's roughly $375 (approx.); at $100,000, roughly $2,179 against roughly $1,375 (approx.). The two lines converge as income rises toward the mid-$100,000s and cross at around $146,486 (approx.), the point above which the original, more generous-at-the-top design starts to outrun the redesign. Above that point the gap widens quickly: at $180,000 the original design would have delivered roughly $6,075 against the redesign's roughly $3,729 (approx.), and at $200,000 and above, where both designs flatten out, the original design's $9,075 (approx.) is almost exactly double the redesign's $4,529 (approx.).
It is worth placing the $146,486 crossover point against how Australians actually earn. Average full-time adult ordinary-time earnings sat roughly in the $90,000-to-$100,000 range through the period the redesign was debated (approx.; ABS, Average Weekly Earnings, various releases 2023-2024), and median individual income sits well below the average, pulled down by the large number of part-time and lower-paid workers in the figures (approx.; ABS). On either measure, the crossover point sits comfortably above where most working Australians actually earn. That is the statistical basis for the government's central claim — that every taxpayer still got a cut, and most got a bigger one — a claim that holds up against the primary income-distribution data, even though it says nothing about whether the smaller group above the crossover point was treated fairly against what it had specifically been promised.
The scale of the reversal, in taxpayers and dollars
Treasury's own numbers, released alongside the January 2024 announcement, put the scale of the redistribution in population terms. Of the roughly 13.6 million Australians who pay income tax (approx.; ATO/Treasury, 2023-24), an estimated 11.5 million were better off under the redesign than they would have been under the original 2019 plan, and an estimated 1.2 million — everyone earning above the crossover point — received a smaller cut than they had been legislated, though still a cut, not a tax rise (approx.; Treasury, Jan-2024). Roughly 2.9 million taxpayers earning below $45,000, who would have received nothing at all under the original design, received a genuine cut for the first time under the redesign (approx.; Treasury, Jan-2024). Those figures are the clearest quantitative answer to the "who gained, who gave something up" question this topic opened with: a very large majority gained more, a smaller number — concentrated at the very top of the income distribution — gained less than they had been promised.
The critique: an explicit promise, broken in plain sight
Start with what makes this different from most political backflips. Most cost-of-living pledges are deliberately vague — "easing the pressure," "help is on the way" — precisely because vague language is hard to hold a government to later. The Stage 3 commitment was the opposite of vague. Ministers were asked directly, on the record, on multiple occasions across more than eighteen months, whether the government would change the design, and the answer was consistently no, with no caveat attached about economic conditions, no hedge about "provided nothing changes." A promise stated that plainly, that often, is not political colour. It is a fact voters, employers and financial planners were entitled to build around.
The reversal was also not triggered by new information about the policy itself. The distributional skew in the original design — nothing below $45,000, a flat maximum benefit above $200,000 — was visible, and was criticised, from the moment the 2019 Act passed. Nothing about the bracket structure was a discovery made in January 2024; the mathematics had not changed since 2019. What had changed was the size of the cost-of-living crisis and the political cost of a design that looked increasingly indefensible against it. That distinction matters, because it separates this case from an ordinary broken promise forced by an unforeseeable external shock — the kind this masthead documents elsewhere, such as the $275 energy pledge overtaken by a war nobody predicted. Here, the government reversed a specific commitment about a policy whose flaws it had always understood, once the political conditions for reversing it looked more favourable than the political conditions for keeping it.
Give the process its due, because it is a genuine point in the government's favour. The reversal was not buried or disguised. Albanese announced it in a set-piece address, at the National Press Club, took questions on it, and released Treasury's modelling and distributional tables into the public domain — in more transparent detail than the original 2019 package ever received. The bill went through ordinary parliamentary scrutiny rather than being rushed through under emergency powers. Whatever the verdict on the earlier commitment, this was an open reversal, argued for in public, not a quiet redrafting nobody was meant to notice.
The "everyone still gains" defence is true but incomplete. Roughly 1.2 million taxpayers earning above the crossover point were told, specifically and by name, that the legislated cut coming their way would not shrink. It shrank — from $9,075 to $4,529 a year at $200,000 and above (approx.), a reduction of roughly half. That is a real, measurable loss relative to what was promised, even though it is not a loss relative to the law that applied before Stage 3 existed at all. A promise is not kept simply because the person it was made to ends up better off than they would have been with no policy whatsoever; it is kept, or broken, against the specific terms of the promise itself.
The practical cost of that specific broken promise showed up away from Canberra, too. Mortgage brokers, financial advisers and household budgeting tools had spent much of 2022 and 2023 modelling clients' post-2024 take-home pay on the original, larger Stage 3 numbers, because that was the law and the government had said, repeatedly, that it would stay the law. Some higher-income households had structured salary packaging, voluntary superannuation contributions or simply their own mortgage-serviceability assumptions around a tax cut that, for many of them, arrived at roughly half the size they had been told to expect. None of this amounts to hardship on the scale of the cost-of-living pressures documented elsewhere on this site — everybody above the crossover point remained, in absolute terms, better off than under the pre-Stage-3 law. But it illustrates a cost of promise-breaking that a purely aggregate count of winners and losers misses: individual financial decisions, made in good faith on the strength of a specific government commitment, that had to be quietly unwound once the commitment changed.
There is also a narrower economic critique of the redesign's timing, separate from the trust question. In January 2024, the Reserve Bank was still working to bring inflation back within its target band, having lifted the cash rate through the most aggressive tightening cycle in a generation (approx.; RBA). Lower- and middle-income households generally spend a larger share of any extra dollar than higher-income households, who save more of it — a well-established pattern in household consumption data. By shifting a larger share of the tax cut toward the households most likely to spend it quickly, the redesign arguably added a little more near-term demand into an economy the Reserve Bank was still actively trying to cool, compared with a design that sent more of the same money to higher earners with a lower propensity to spend it immediately. Treasury and the government argued the difference was too small to matter for inflation, and that cost-of-living relief for stretched households was the more urgent priority; not every economist agreed on how to weigh those two considerations against each other (approx.; RBA and Treasury commentary, 2024). The trust question and this narrower demand-timing question are both legitimate lines of critique, and they do not point in exactly the same direction about what "good policy" meant in this specific instance.
A promise stated that plainly, that often, is not political colour. It is a fact voters were entitled to build around.
This is also not the first time a promised tax outcome has not survived contact with the political cycle. The Low and Middle Income Tax Offset, a temporary measure worth up to $1,500 a year for many middle earners, was allowed to lapse after the 2021-22 financial year under the outgoing Coalition government's own 2022-23 Budget, and the incoming Labor government did not reinstate it — delivering what amounted to a real cut in after-tax income for millions of workers just as the cost-of-living crisis was beginning to bite (approx.; Budget Paper No. 2, 2022-23, Mar-2022). That history does not excuse the Stage 3 reversal. It does show that tax relief promised for one year, or one term, is routinely treated by governments of both colours as more negotiable in practice than it is presented as being at the time it is announced.
Put plainly: both readings of this episode survive contact with the evidence, and they are not in tension with each other. The redesign was, on the numbers, better-targeted policy for the moment it landed in — more money to more people, delivered at no extra cost to the budget, when the cost-of-living crisis was at its worst. And delivering it required breaking an explicit, specific, repeatedly reaffirmed commitment that the government itself had chosen to make in absolute terms, long after it had every opportunity to qualify it. Good policy and a broken promise are not mutually exclusive outcomes here. They are simply two different measures of the same decision, and a masthead built on checking the record against the promise has an obligation to report both.
What the government got right
Credit belongs where the evidence supports it. The redesign fixed a real and long-standing design flaw — a tax cut worth nothing below $45,000 and worth its maximum only above $200,000 — that had been criticised on distributional grounds since before Labor took office, including by economists who were never natural allies of this government's broader agenda. It did so without adding to the cost of the package relative to the original plan, meaning the improvement in targeting was not funded by extra borrowing or an offsetting tax rise elsewhere. And it was delivered in the year the cost-of-living crisis was doing the most damage to middle and lower earners, rather than being shelved for a more politically convenient moment closer to an election.
The government also did not try to hide the reversal behind technical language or a quiet regulatory change. It was announced at a set-piece address, defended in interviews, and passed through ordinary parliamentary process with published costings and distributional tables that let anyone — a journalist, an economist, an individual taxpayer — check exactly how their own income band was affected. That level of transparency is not guaranteed in tax policy, and it is fair to note the original 2019 package, when it passed, received less granular public modelling of its own distributional effects than this reversal did.
It is also worth restating what the redesign did not do. It did not raise anyone's tax bill relative to the law that applied before Stage 3 was ever legislated. Every taxpayer who stood to gain from the original design still gained from the redesign — some gained more than promised, a smaller group gained less than promised, but nobody was left worse off than if Stage 3 had never existed at all. That is a materially different outcome from a government that funds a popular new policy by quietly clawing money back from a group that had been told it was safe.
The reception among independent economists and analysts was, on balance, more positive than negative. Several commentators with no particular loyalty to either major party argued at the time that the redesign was simply better tax design than the plan it replaced — more progressive, no less fiscally responsible, and correcting a flaw that arguably should never have been legislated in the flat, single-bracket form it took in 2019 (approx.; commentary from independent economists and research organisations, Jan–Feb 2024). That reception does not settle the trust question this topic is built around. But it is relevant to the other half of the question: measured purely as tax design, independent of the process that delivered it, the redesign attracted meaningfully less expert criticism than the plan it replaced.
Back at the service desk, none of this changes what actually happened to Renata Cho's pay. The redesign put money in her pocket that the original, more generous-at-the-top design never would have. Told honestly, her story is a genuine data point in the government's favour: a low-income worker, previously excluded entirely from a five-year-old tax plan, finally included in one, at the exact moment the cost-of-living crisis was squeezing her hardest. What she was not part of, and could not have been, is the separate question this topic exists to answer — whether a government that spent almost two years telling a different group of taxpayers, in absolute and unqualified terms, that their legislated cut was safe, was entitled to break that specific word once a fairer design became available. Both things are true at once. That is not a contradiction. It is simply what an honest account of this policy looks like.
The risk to the future
If this promise wasn't safe, which one is?
The lasting risk here is not the dollar figures — it is the precedent. A specific, legislated, multi-year tax structure, reaffirmed in unconditional terms dozens of times over almost two years, was rewritten anyway once it became politically convenient to do so. Australia is now on notice that "it's already legislated" and "we've made a commitment" do not, by themselves, guarantee that a tax setting will survive to the date it takes effect. That matters well beyond Stage 3. The country is currently relying on a string of similarly load-bearing, multi-year commitments — the trajectory of the $3 million superannuation tax threshold examined elsewhere on this site, the indexation (or non-indexation) of income tax brackets, the long-run schedule of the superannuation guarantee, defence spending commitments stretching into the 2030s — each of which depends, for its credibility with households, businesses and financial planners, on the assumption that a legislated, repeatedly-reaffirmed setting is actually fixed.
None of this means governments should never revisit a policy when circumstances genuinely change; rigidity for its own sake is not good governance either, and a government that could never adjust a five-year-old tax schedule regardless of the economic conditions around it would be a worse outcome, not a better one. The risk is narrower and more specific: when a promise is made as absolutely and as repeatedly as this one was, and is broken anyway, the next government's promise — on tax, on superannuation, on any multi-year commitment this masthead documents — is worth a little less to the people who have to plan their lives around it. That erosion compounds. A country that cannot trust its Parliament's legislated word on tax settings will eventually have to price that uncertainty into every major financial decision its households and businesses make, from a mortgage taken out on an assumed take-home income to a retirement plan built around a superannuation schedule that may or may not still exist when it matters.