Consider a paramedic in outer Melbourne on a base salary of around $85,000 a year (approx.) — the kind of income that used to sit comfortably in the middle of the pack. Enterprise bargaining delivers her a pay rise most years, usually pitched somewhere near inflation, just enough to stop her falling behind at the checkout and the bowser. On paper the rise looks like progress: a bigger number on the payslip, a small win against the cost of living. But nothing in the tax scale moved to meet her. The thresholds that decide how much of her next dollar the Australian Taxation Office keeps are the same dollar figures they were five years ago, or ten. Each pay rise, however modest, pushes a larger share of her income into the 30 per cent bracket rather than the 16 per cent one below it. She has not become richer in any way she would recognise at the shops. She has simply crossed a line that Parliament drew years ago and never got around to redrawing.
Multiply her by the roughly 14 million (approx.) Australians who lodge an individual tax return each year, and the quiet arithmetic becomes one of the largest sources of new government revenue in the country — bigger, in dollar terms, than most single policy announcements a treasurer will ever make. It has a name: bracket creep, sometimes called fiscal drag. It needs no legislation, no question time, no vote. It is the default setting of an income tax system whose brackets are fixed in nominal dollars while wages and prices are not. Governments of both major parties have, every few years, legislated a tax cut to hand some of it back. The question this piece sets out to answer is how much of that giveback actually survives contact with the next few years of inflation — and who is left holding the bill for the part that does not.
What the government did: tax cuts that reset a scale it never re-drew
Australia's personal income tax scale has a tax-free threshold and marginal rates that step up as income rises above it. For the 2024-25 income year, following the redesign of the already-legislated Stage 3 cuts, the scale runs: no tax on the first $18,200; 16 per cent on the next slice up to $45,000; 30 per cent from $45,001 to $135,000; 37 per cent from $135,001 to $190,000; and 45 per cent above that (Australian Taxation Office, individual income tax rates, 2024-25). None of those dollar thresholds is written into law with any instruction to move automatically with inflation or wages. They sit in the tax law as fixed numbers, the way they always have, and they stay exactly where Parliament last left them until Parliament passes a new law to shift them.
That matters most at the bottom of the scale. The tax-free threshold — the amount every taxpayer earns before the ATO takes a cent — was last raised in the 2012-13 income year, when it jumped from $6,000 to $18,200 as part of the household assistance package that accompanied the carbon price (Australian Taxation Office, historical individual tax rates). It has not moved since. More than a decade of cumulative inflation and wages growth has passed under a line that has not shifted by a single dollar.
Government has not ignored the problem; it has addressed it periodically, through discretionary cuts rather than a standing indexation rule, and Stage 3 was not the first attempt. An earlier three-stage plan, legislated by the Coalition government in 2018, had already delivered two rounds of bracket-creep relief before Stage 3 existed. Stage 1 introduced the Low and Middle Income Tax Offset (LMITO), a temporary, means-tested rebate paid at tax time rather than a permanent change to the rate scale. Stage 2, brought forward to apply from the 2020-21 income year, lifted the threshold for the 32.5 per cent rate from $37,000 to $45,000 and the threshold for the 37 per cent rate from $90,000 to $120,000 (Australian Taxation Office, historical individual tax rates; Treasury Laws Amendment (Personal Income Tax Plan) Act 2018-2020, approx.). LMITO was never intended to be permanent, and it lapsed after the 2021-22 income year — meaning millions of low- and middle-income taxpayers who had grown used to an offset worth up to $1,500 (approx.) at tax time simply stopped receiving it, at the same moment bracket creep on the un-indexed core scale kept pushing their average rate up regardless. The sequence illustrates the underlying pattern well before Stage 3 or the 2025-26 Budget cuts existed: relief arrives in discrete parliamentary decisions, some of it temporary by design, while the mechanism generating the need for relief runs continuously in the background no matter which relief measure happens to be in place at any given moment.
The most recent major reset to the permanent rate scale was the January 2024 redesign of the Stage 3 tax cuts — legislated by the Morrison government in 2018, reshaped by the Albanese government before taking effect on 1 July 2024 — which lowered the second-bracket rate from 19 to 16 per cent, cut the 32.5 per cent rate to 30 per cent, and lifted the thresholds at which the 37 and 45 per cent rates begin (Treasury Laws Amendment (Cost of Living Tax Cuts) Act 2024; Australian Taxation Office, 2024-25 individual tax rates). That redesign, examined in full in a separate audit on this site, changed who benefited from the legislated cut more than it changed its total cost, and it is discussed here only for what it did to the shape of the bracket scale.
A second, smaller reset followed just over a year later. In the 2025-26 Budget, handed down in March 2025, the government announced a further cut to the rate above the tax-free threshold: 16 per cent falling to 15 per cent from 1 July 2026, and to 14 per cent from 1 July 2027 (Australian Government, Budget Paper No. 2, March 2025, approx.; Australian Taxation Office, "Personal income tax — new tax cuts for every Australian taxpayer"). The government's own material described the cuts explicitly in terms of returning bracket creep, promising every taxpayer up to $268 more a year from the first stage and up to $536 a year, cumulatively, from the second, each measured against 2024-25 settings (approx.; ATO, personal income tax new cuts explainer, 2025). Both changes are real, legislated relief. Both are also, by design, one-off resets of a scale that starts creeping again the moment they take effect, because nothing about how thresholds are set going forward actually changed.
The strongest case for the policy
Every un-indexed income tax system works this way — and periodic cuts let Parliament target the relief
Put the government's defence at its best, and it has three parts. First, Australia is not unusual: fixing dollar thresholds and periodically resetting them by legislation, rather than indexing them automatically, is close to the historical norm for this country's income tax system, under governments of both major parties, going back decades. Australia has already tried automatic indexation once and abandoned it: following the 1975 Mathews Inquiry into Taxation and Inflation, the Fraser government indexed tax brackets to inflation from the 1976-77 income year, moved to only half-indexation the following year, and dropped automatic indexation altogether by 1981-82 (approx.; Parliamentary Budget Office, tax-mix history, "1980 to 2000: Tax reform," approx.). Indexation locks in the current bracket structure and its distributional pattern indefinitely — itself a policy choice a future Parliament might reasonably want to revisit rather than have fixed by formula. Every tax cut since — the Costello-era cuts of the 2000s, the Coalition's three-stage plan legislated in 2018 and 2019, and Labor's redesign of Stage 3 plus the further cuts announced in the 2025-26 Budget — has been a live decision, debated in Parliament and defended at an election, about who should get the benefit and by how much.
Second, that live decision-making lets a government target relief in a way a mechanical indexation formula cannot. The clearest example sits inside this very story: the redesign of Stage 3 explicitly redirected more of the benefit toward low- and middle-income earners and away from the top of the scale, a deliberate progressivity choice that a formula tying every threshold to CPI or wages growth would not have made on its own. Indexation is blind to distribution; discretionary reform is not. Third, on the government's own account, the cuts it has now legislated — Stage 3 from 2024-25, and the further 16-to-15-to-14 per cent cuts from 2026-27 and 2027-28 — are explicitly framed and marketed as returning bracket creep, not merely delivering a one-off saving. A reasonable observer can conclude the system, imperfect as it is, is doing roughly what it is meant to do: letting inflation do some quiet, unlegislated revenue-raising between elections, then having an elected Parliament hand a meaningful share of it back through a debated, targeted, and — on the numbers — genuinely substantial tax cut.
The evidence: what the primary record shows about the trend, not just the cuts
The primary record does not dispute that periodic cuts return real money. It disputes how much of the underlying trend they change. Three sources tell the same story from different angles: the Australian Taxation Office's own rate schedules going back over a decade, showing how far the frozen tax-free threshold has drifted below what inflation alone would suggest it should be; the Parliamentary Budget Office's running analysis of the average personal income tax rate — the share of every dollar earned that ends up paid in tax, once averaged across all the rates a person's income passes through; and the same office's projections of personal income tax as a share of total Commonwealth tax receipts, the clearest gauge of how reliant the federal budget has become on this one, largely automatic, revenue stream.
A threshold frozen while everything around it moved
The clearest illustration sits at the very bottom of the scale. The $18,200 tax-free threshold has applied, unchanged, to every income year since 2012-13 (Australian Taxation Office, historical individual tax rates). Consumer prices have risen by a little under half over that period, on the Australian Bureau of Statistics' Consumer Price Index (ABS, Consumer Price Index, Australia, various releases, 2012-2025, approx.). A threshold that moved in step with prices alone — leaving aside wages, which have generally grown faster than prices over the same stretch — would likely sit somewhere in the mid-$20,000s today rather than at $18,200 (approx.; derived from ABS CPI series). Every dollar of income between the frozen threshold and where an indexed one would now sit is a dollar that used to be tax-free and no longer is, for no reason connected to any vote Parliament has taken on the tax-free threshold itself since 2012.
The one time Australia tried automatic indexation
The historical record contains one direct test of the alternative. Following the 1975 Mathews Inquiry into Taxation and Inflation, the Fraser government began indexing tax brackets automatically to inflation from the 1976-77 income year. Full indexation lasted a single year before the government moved to only half-indexation, and automatic indexation was dropped altogether by 1981-82 (approx.; Parliamentary Budget Office, tax-mix history, "1980 to 2000: Tax reform," approx.). Budgetary pressure was the main reason it was abandoned: an indexed scale collects less revenue by design than an un-indexed one facing the same inflation, and a government that had committed to indexation found itself needing more revenue than the formula was delivering. Contemporary accounts of the decision also note a second, more uncomfortable reason: bracket creep is less visible to voters than an explicit tax increase, which made reversing indexation politically easier than raising a headline rate by the same amount (approx.; Parliamentary Budget Office, tax-mix history, approx.). Every government since has governed under the system that decision restored — fixed thresholds, periodic discretionary relief — rather than the indexed one it briefly replaced.
The average tax rate, and where it is headed
The Parliamentary Budget Office tracks a single number that captures the cumulative effect better than any bracket-by-bracket comparison: the average personal income tax rate, meaning total personal income tax collected as a share of total taxable income across the economy. Over roughly the past forty years that rate has moved in a fairly narrow band, mostly between 22 and 25 per cent, rising with bracket creep between tax cuts and falling back when a cut lands (Parliamentary Budget Office, bracket creep and tax-mix analysis, various, 2021-2024, approx.). The PBO put the current rate at around 24.9 per cent, and — in scenario work assuming no further tax cuts are legislated beyond what is already on the books — projected it climbing to around 28.5 per cent by the early-to-mid 2030s, a level the office's own commentary describes as unusually high by the standard of the past forty years (Parliamentary Budget Office, bracket creep and tax-mix scenario analysis, approx. 2024). Some later PBO analysis goes further still: once the office also factors in the slow structural decline of fuel and tobacco excise — two revenue lines shrinking as vehicles electrify and smoking rates fall — its modelling has the average income tax rate pushing above 30 per cent, as those other revenue sources are effectively replaced by more tax on wages (Parliamentary Budget Office, cited in contemporary reporting on the office's fiscal analysis, approx. 2024).
The pattern is not a straight line; it saws up and down with every tax cut, then resumes climbing the moment the cut takes effect, because nothing about the underlying mechanism changed. Research using the long-running HILDA household survey found the average tax rate paid by full-time workers rose from around 18.1 per cent to around 20.3 per cent between the 2021-22 and 2022-23 tax years alone (approx.; HILDA Survey-based analysis, University of Melbourne, reported 2025-2026) — a jump driven mainly by the lapsing of the temporary Low and Middle Income Tax Offset after the 2021-22 year, rather than bracket creep on its own, but a clear illustration of how quickly the average rate can move once temporary relief runs out and thresholds stay still. The same research found the Stage 3 redesign pulled the full-time average rate back down by around two percentage points in 2024-25 (approx.) — a real and immediate saving, before the same climbing pattern resumes underneath it.
Personal income tax's growing share of everything Canberra collects
The second measure concerns the shape of the whole tax system rather than any one taxpayer's rate. The PBO estimates personal income tax currently accounts for somewhere in the order of 40 to 42 per cent of total Commonwealth tax receipts (approx.; Parliamentary Budget Office, 2024). Its scenario work — built across different report vintages and modelling assumptions, so best read as a consistent direction rather than one precise number — has that share climbing toward somewhere between the high 40s and mid-50s per cent of total receipts by the early-to-mid 2030s if bracket creep runs unchecked (approx.; Parliamentary Budget Office, various tax-mix and Medium-Term Budget Outlook publications, 2021-2026). Even the more conservative end of that range would put personal income tax's share close to or above the roughly 47 per cent (approx.) peak Australia last saw in the mid-1980s, before the GST broadened the revenue base away from income tax (Parliamentary Budget Office, historical tax-mix analysis, approx.). The PBO's most recent Medium-Term Budget Outlook work, reported in mid-2026, went further: it identified bracket creep as the single largest driver of the budget's own projected return to surplus around 2034-35, and warned that relying on it that way assumes government holds off legislating any further income tax relief for roughly a decade — an assumption the office itself called unrealistic, not least because the Treasurer has already flagged further relief (Parliamentary Budget Office, Medium-Term Budget Outlook, 2026, approx.).
Every tax cut this decade has been sold as returning bracket creep. None of them has been asked to stop it from starting again the next morning.
The critique: a tax rise that needs no vote
Bracket creep's defining feature is procedural, not just fiscal: it raises the government's revenue take from wages every single year, automatically, without a minister ever standing at a despatch box to argue for it and without a single MP casting a vote for it. Compare that with the last time Australia changed its consumption tax rate — the introduction of the GST in 2000 required an election campaign, a Senate negotiation, and a specific piece of legislation subject to full parliamentary scrutiny. A rise in the average income tax rate driven by bracket creep requires none of that. It happens by default, and it only stops or reverses when a government chooses to spend political capital legislating a cut. That asymmetry — automatic increase, discretionary decrease — is precisely why the average rate saws upward between cuts rather than sitting flat: the default direction of travel is up, and reversing it is always optional, for whichever government happens to hold office when the next budget is written.
Every year Canberra doesn't legislate a change to the tax scale is a year the tax system changes anyway.
Australia's approach is not universal, and it is not the only approach the Australian tax system itself uses. Comparative analysis of public finances across advanced economies has found that a substantial minority of OECD members adjust income tax brackets for inflation automatically each year, with the United States among the most prominent examples — its federal tax brackets, standard deduction and a range of other thresholds have moved with the Consumer Price Index annually since the mid-1980s, without requiring a fresh act of Congress each time (approx.; IMF, "Inflation Indexation in Public Finances: A Global Dataset," working paper, 2023). Australia's own Medicare levy is a small proof of the concept working close to home: the income thresholds at which the levy applies, and at which low-income earners are exempt or pay a reduced rate, are indexed annually to growth in Average Weekly Ordinary Time Earnings, a wages-based measure the Australian Bureau of Statistics publishes each year (Australian Taxation Office, Medicare levy low-income thresholds, updated annually). The same government that leaves the core income tax scale fixed adjusts a companion levy's thresholds every single year. That is not proof indexing the main scale would be costless — the earlier section on what the government got right addresses that trade-off directly — but it is evidence that the absence of indexation on the scale that matters most is a policy choice, not an administrative impossibility.
Bracket creep does not fall evenly. Because Australia's scale is progressive — each extra dollar taxed at a higher rate than the last — the proportional bite of an un-indexed threshold is largest for people who have only recently crossed into a higher bracket, typically low- and middle-income earners moving out of the tax-free zone into the 16 per cent bracket, or out of it into the 30 per cent one. The PBO's own analysis has found that, over recent decades, the share of total personal income tax paid by the top 10 per cent of earners has fallen — from well over 60 per cent in the early 1950s to below 40 per cent by the 1980s, and on a trajectory to keep falling further by the early 2030s (approx.; Parliamentary Budget Office, trends in personal income tax, approx. 2023) — not because the wealthy are paying less tax in absolute terms, but because bracket creep pulls more and more low- and middle-income taxpayers up the scale toward them. A tax system that becomes more reliant on bracket creep is, by construction, a tax system quietly becoming less progressive over time, even while every headline rate on the books stays exactly as legislated.
The same automatic quality that makes bracket creep politically convenient makes it fiscally consequential. Elsewhere on this site, an audit of the "spending ratchet" describes five areas of Commonwealth spending — the National Disability Insurance Scheme, aged care, health, defence and the interest bill on accumulated debt — growing faster than the economy in a way that is structural rather than cyclical. Every one of those pressures needs to be funded from somewhere. Broadening the GST base, introducing a land tax, or taxing capital income more like wages are all live ideas raised periodically by tax reviews going back to the 2010 Henry Tax Review, and all of them require a government to spend scarce political capital defending an explicit, visible tax rise. Bracket creep requires none of that. It simply happens, quietly, in the background, while Parliament argues about other things — and it is, on the PBO's own numbers, doing more to close the gap between what government spends and what it raises than any single deliberate tax reform legislated this decade.
The burden also falls unevenly by the type of income involved, not just its size. Wage and salary income is captured through Pay As You Go withholding in real time, every pay cycle, with essentially no scope for a worker to defer, restructure or shelter it from bracket creep's effect. Income from capital — superannuation earnings taxed at a flat, concessional rate, the capital gains tax discount, negatively geared investment losses offset against other income — is not exposed to the same automatic, threshold-driven mechanism in anything like the same way. A tax system leaning more heavily on bracket creep each year is, mechanically, a tax system leaning more heavily on the income of people who cannot easily restructure around it: employees more than investors, and earlier-career workers crossing thresholds for the first time more than retirees drawing largely tax-concessional superannuation income.
What they got right, and the caveats
Credit is due in specific places. The government did legislate real relief, twice, in a single term — the Stage 3 redesign from 1 July 2024, and the further 16-to-15-to-14 per cent cuts from 1 July 2026 and 1 July 2027 — and both were explicitly marketed and defended in Parliament as a response to bracket creep, not disguised as something else. The scale of relief is not trivial: Treasury's own costings put the cumulative value of the legislated Stage 3 changes, relative to the pre-2024 scale, in the order of $250 billion (approx.) over the decade to the early 2030s (Australian Government, Budget Paper No. 2 and related costings, approx. 2024), and the further 2025-26 Budget cuts add a genuine, if smaller, annual saving on top for every taxpayer above the tax-free threshold. A taxpayer who banks both rounds of relief is unambiguously better off in nominal terms than they would have been under the pre-2024 scale.
It is also fair, in the interest of the fairness this site holds itself to, to note that reliance on bracket creep long predates this government and is not a partisan invention. The Fraser Coalition government's own experiment with automatic indexation, run from 1976-77, was wound back by that same government within five years, and the nine years of Coalition government between 2013 and 2022 saw their own long stretches without a legislated tax cut, during which bracket creep did exactly the same quiet work it is doing now. Every cut since — Coalition and Labor alike — has been explicitly framed, at least in part, as returning some of what bracket creep took. The pattern examined here is a structural feature of how Australia's income tax has been run for at least two generations, under governments of both colours, not a mechanism this government invented or could unilaterally end without a scale of tax reform no government of recent decades has attempted.
Some credit also belongs to the transparency the system does provide, even without an indexation mechanism. The Parliamentary Budget Office was established by legislation passed in December 2011 and began operating the following year, specifically to give Parliament and the public independent costings and projections from outside the Treasury and Finance departments that answer to the government of the day, and its recurring bracket-creep analysis — the same analysis this audit draws on throughout — is a large part of why the mechanism is visible and debatable at all, rather than a hidden technical drift no outside body tracks (Parliamentary Service Amendment (Parliamentary Budget Officer) Act 2011, Commonwealth of Australia). A government that funds and retains an independent office willing to publish projections showing its own growing reliance on an unlegislated revenue source is, at minimum, not actively concealing the trend, even if it has not yet legislated a permanent fix for it.
It is worth stating plainly, too, that automatic indexation is not a costless fix. Every dollar bracket creep currently delivers to the budget is a dollar a fully indexed system would not automatically collect, and would therefore need to be found some other way — through a broader consumption or land tax base, slower spending growth, or accepting a structurally larger deficit. None of the major reviews that have examined the idea, including Treasury's 2015 Tax White Paper process, recommended full automatic indexation without pairing it with some other significant, and politically difficult, change to the revenue base. The debate is not between bracket creep and a free lunch; it is between one revenue-raising mechanism that requires no legislative act, and several others that all require one.
The risk to the future
A tax system leaning harder on wages precisely when spending is growing hardest
The forward-looking risk is not that bracket creep exists — it has existed, in some form, for as long as Australia has had a progressive income tax with fixed dollar thresholds. The risk is the trajectory the PBO itself has now put a number on: an average personal income tax rate climbing from a current level already near the top of its forty-year range toward the high 20s, or above 30 per cent once other revenue lines are accounted for, and a personal income tax share of total Commonwealth receipts drifting back toward, or past, levels not seen since before the GST existed — all without a single new tax having been legislated. That is a tax rise, in every sense that matters to the person paying it, delivered by default rather than by debate.
It compounds directly with the structural spending pressures documented elsewhere on this site. The NDIS, aged care, health, defence and the interest bill are all growing faster than the economy for reasons that will not resolve themselves within a single budget cycle, and the PBO's own analysis identifies bracket creep as the single largest force currently projected to close that gap — not tax reform, not spending restraint, not a broader revenue base, but the quiet mechanical effect of un-indexed thresholds meeting an inflating, growing economy. A government that lets that mechanism do the heavy lifting for a decade is choosing, whether it says so explicitly or not, to fund the growing size of the state disproportionately from the wages of people still working and still climbing the income scale — precisely the cohort least able to shelter income the way capital owners and retirees more easily can, and the cohort with the most working years still ahead of it to feel the effect compound.
The next generation inherits two versions of that choice. If bracket creep continues to be only partially and periodically returned, today's younger and middle-income workers will spend a larger share of their working lives paying a materially higher average tax rate on their wages than their parents did at the same age and income, before any new tax is ever legislated against them by name. If, instead, a future government chooses to break the cycle with full indexation, it inherits the harder and more honest task every review of the system has pointed to: finding a durable, broader-based way to fund an ageing, more expensive state, rather than letting inflation quietly do it in Parliament's absence. Either path is a legitimate choice for an elected government to make. What is not legitimate, on the evidence assembled here, is continuing to make it by default, one un-indexed dollar at a time, without ever being asked to defend it as one of the largest tax increases most working Australians experience in any given decade.