Passed. Not yet permanent.

By James Zhang · Tax · Updated 13 September 2026 · 31 min read

A law with an election standing between it and its own start date

Most tax changes are settled the day they receive assent. This one is not, and the reason is a scheduling accident that nobody designed.

The Treasury Laws Amendment (Tax Reform No. 1) Act 2026 received Royal Assent on 26 June 2026. Both of its headline measures — quarantining negative gearing on established residential property, and replacing the 50% capital gains discount with cost-base indexation plus a 30% minimum tax — commence on 1 July 2027. Under the Constitution, the writs for a half-Senate election cannot be issued before 1 July 2027 either, which puts the earliest possible federal election on Saturday 7 August 2027 and the latest on Saturday 20 May 2028.

So the sequence is: legislate, wait fourteen months, start collecting, and then face the voters somewhere between five weeks and eleven months later. There has rarely been a cleaner set-up for a repeal campaign, and the Opposition has already committed to one in terms that leave no wriggle room.

And yet the probability that any of it is actually repealed is low — on our reckoning, about one chance in ten, with the odds that a post-May-2026 buyer ever gets their negative gearing back closer to one in thirty. The reason has almost nothing to do with whether the policy is good, and almost everything to do with a Senate, a fractured centre-right primary vote, and a design feature that gets more expensive to unwind every month it sits on the books.

This page sets out what actually has to happen, in order, and puts a number on each step. It also deals honestly with the part of the government’s case that has aged worst — the modelled $2 a week — and with the part of the investor case that does not survive contact with the polling.

What was actually legislated, and what was not

A great deal of the commentary since May has argued against a policy that was not passed. It is worth being precise, because the details are exactly what determine how hard the thing is to unwind.

Negative gearing. From 1 July 2027, losses on established residential property acquired after 7:30pm AEST on 12 May 2026 can only be deducted against residential property income — rent, and capital gains on rental property. Unused losses are carried forward indefinitely rather than lost. Property held at the moment of the announcement is grandfathered and keeps the old treatment for as long as it is held. New builds are exempt entirely.

Capital gains tax. From 1 July 2027, the 50% discount for individuals, trusts and partnerships is replaced with cost-base indexation plus a minimum 30% tax rate on the gain. This applies to gains accruing after 1 July 2027 only, and — this is the part most often mis-stated — it applies to all CGT assets, not just property. Nothing is deemed to be sold on 1 July 2027 and no tax falls due on that date; the split between pre- and post-2027 growth is worked out later, at realisation, by valuation or a specified apportionment formula. Main residences remain exempt. Pensioners and income support recipients are exempt from the 30% minimum. New-build investors may elect the old 50% discount instead.

Two carve-outs were conceded before the bill passed. On 18 June 2026 the 50% discount was reinstated for startups and small businesses with turnover under $10 million. On 23 June the Greens traded their votes for an end to the SMSF limited-recourse borrowing exemption on residential property, and for concessions on the NDIS inquiry. The bill passed the Senate on 25 June with Greens support.

It is not finished. A second tranche of legislation — defining a “new residential dwelling”, setting the gain and loss apportionment rules, and dealing with joint tenancy, separation and inheritance — went to consultation closing 21 August 2026, with further tranches expected. A scheme still being drafted is a scheme with a running target painted on it, and the Opposition has been shooting at it accurately.

The last time Australia tried this

Australia has run this experiment once. It lasted twenty-four months, it was reversed by the government that introduced it, and almost everything said about it in public is wrong in one direction or the other.

On 17 July 1985 the Hawke government quarantined rental property losses: they could no longer be deducted against wages or other income, only against rental income, with the excess carried forward. Farm losses were quarantined at the same time. On 15 September 1987, in the Budget, it was reversed, backdated to 1 July 1987.

Two features genuinely do differ, and they are the ones worth arguing about. First, lead time: 1985 took effect on the day it was announced; 2026 gives fourteen months' notice. Second, and far more important, the new-build exemption. In 1985 the quarantine applied to new and established property alike, offset only by a 4% building depreciation allowance introduced at the same time — which was then cut back to 2.5% when negative gearing was restored, so the new-build sweetener was traded away for the return of general gearing. In 2026 new construction is exempt outright and keeps the 50% CGT discount as well. The supply channel that the entire industry argument runs through has been deliberately left open.

What actually happened to rents

This is where both sides quote true numbers from different columns, and where the argument has been stuck for forty years.

Nominal rents rose in every capital city. That is true, it is what people remember, and the reform side almost never says it out loud. But inflation in those years ran at 8–9%. The CPI rose 18.6% between June 1985 and June 1987. A 16% nominal rent rise over that window is a 2% real fall.

In real terms, rents rose materially in exactly two cities.

Saul Eslake’s argument from this is the strongest one on the reform side, and it is structural rather than statistical: if quarantining had triggered a landlords’ strike, rents should have risen everywhere. They rose fastest in the two cities with the tightest vacancy rates, and fell in real terms in four others.

Keating’s own cabinet submission in 1987 said the same thing, which is remarkable given his government reversed the policy weeks later: with the exception of Sydney, rental conditions were “not unusually tight,” and “the evidence suggests that local influences, rather than tax measures, dominate.”

The reform side’s case has three real weaknesses, and it is worth naming them rather than pretending otherwise. The vacancy-rate explanation rests on figures its own advocates state inconsistently — Sydney is variously put at “barely above 1%” and “just below 2%” by people making the same argument. Dwelling completions genuinely did fall, and the timing is at least consistent with the industry story. And most importantly: there is no econometric study of the 1985–87 episode at all. Forty years on, the entire natural experiment has been eyeballed. Nobody has run a difference-in-differences across the eight cities. Every claim on both sides is descriptive.

The supply question, and the confounders that swamp it

Completions fell 16.6% from 143,433 in 1984–85 to 119,633 in 1987–88, then rebounded 27% the year after restoration. Read on its own that looks decisive. It is not, because the 1985 quarantine was never a clean experiment. Running simultaneously were: a brand-new capital gains tax from 20 September 1985, dividend imputation, fringe benefits tax, the removal of the ceiling on owner-occupier mortgage rates in April 1986 — the price of a mortgage was deregulated in the middle of the experiment — mortgage rates moving from about 11.5% to 15.5%, and an All Ordinaries index that doubled between January 1985 and January 1987 before falling roughly 25% in a single day in October 1987.

That last one is decisive and it works in both directions. The equity boom pulled capital out of residential property while the quarantine was in force; the crash pushed it back in two months after negative gearing was restored. Attributing the 1988–89 construction rebound to the tax change alone is not sustainable.

The part of 1985 that should worry a defender of the policy

Not the rents. The politics.

The 1985 measure was not repealed by an incoming government. It was repealed by the government that wrote it, inside two years, without an election intervening. The official reasons were tax-treatment consistency and the argument that the new CGT already dealt with the high-income benefit. The unofficial ones, per the tax literature, were an impending federal election and complaints from a New South Wales government facing a state election of its own. One academic account calls it “one of the more remarkable backflips in Australian tax policy history.”

That is the real lesson of 1985, and it is not the one either camp usually draws. The risk to this policy is not a Coalition repeal bill in 2028. It is a Labor government looking at a Sydney rental market, a state election and its own marginal seats, and deciding fourteen months of notice can quietly become twenty-six. We have moved our probabilities accordingly.

The two gates a repeal has to pass

“We will repeal it” is a sentence about intent. Repeal is a process, and it has to clear three separate obstacles in sequence. Multiply the three and you have the answer.

Gate one: a change of government

This is where most of the probability disappears, and not for the reason investors assume. Labor holds 94 of 150 seats and won the 2025 election 55.2–44.8 after preferences. It has not lost that lead. Across every national poll taken since the budget, Labor’s two-party-preferred figure against the Coalition has sat between roughly 50 and 56, and the September readings have it at 52–48 and 53–47.

What has changed is the shape of the opposition to it. The Coalition’s primary vote has fallen to 18–21%, while One Nation has run at 23–30% and led the Coalition in most polls since June. On 9 May 2026 One Nation won the Farrer by-election with 57.6% two-candidate-preferred in a seat the Liberals had held 56–44 a year earlier — a swing of roughly 31 points, and the first One Nation candidate ever elected to the House of Representatives. In late August it took its first Western Australian lower house seat at the Secret Harbour by-election.

This matters for the repeal question in a way that is easy to miss. A party polling 18% does not form a government on its own. The only arithmetic that produces a non-Labor government is a Coalition–One Nation arrangement, which has never happened federally, which Angus Taylor has declined to rule out, and which senior Liberals have publicly refused to countenance. Andrew Hastie’s line — “I won’t be pushed around by One Nation, nor will I bow the knee to them” — is not a small internal disagreement. It is a description of the fault line that has to close before anyone repeals anything.

Our estimate for gate one: about a 22% probability that the next election produces a government committed to repeal. That is well above what a pure two-party-preferred reading would give you, because it allows for eighteen months of political time, for a leadership change (Taylor’s own position is contested, with Hastie the named alternative), and for the possibility that a fragmented right consolidates behind whoever is ahead.

Gate two: the Senate

This is the gate that almost nobody prices, and it is the one that killed more tax reversals than the ballot box ever has.

A repeal bill needs 39 votes in a Senate of 76. Labor holds 30 today; the Greens are the largest crossbench bloc at around ten. Together they are at or above the line. Only half the Senate faces the voters at an ordinary election, so even a decisive change of government in the House leaves most of the current Senate untouched for another three years. The Greens have said the reforms did not go far enough — Larissa Waters’ complaint was that grandfathering “baked the inequality in” — which means they are the least likely crossbench votes in the building for a repeal.

The counter-case is real: if the right did well enough to take the House, it would likely also do well in the Senate, and One Nation — whose leader called the changes “communism” — would vote for repeal. That is why we put this gate near a coin toss rather than lower.

Our estimate for gate two: about a 50% probability that a repeal-minded government could actually pass a repeal bill in its first term.

Gate three: following through

Governments do repeal taxes they campaigned against. The carbon price was gone by July 2014 and the mining tax by September 2014, both within a year of a change of government elected on an explicit repeal mandate. That precedent is the single strongest argument for a reversal here, and it should not be dismissed.

But it is worth noticing what the incoming government would be walking into. The measure raises $3.63 billion over the forward estimates to 2029–30, and that money is hypothecated in the budget papers to income tax cuts — principally the Working Australians Tax Offset, $250 a year to around 13 million workers from 2027–28. Repealing the revenue without repealing the offset leaves a hole; repealing both means taking $250 a year off thirteen million people in order to restore a deduction claimed by about 230,000 new investors annually. That is not a trade a first-term government makes casually.

Our estimate for gate three: about 85% — high, because a promise this explicit is hard to walk away from, but not 100%, because the fiscal arithmetic gets uglier the longer the measure runs.

Why the polling does not say what investors think it says

Here is the uncomfortable part for anyone hoping for a reversal, and it is worth stating plainly rather than talking around.

The measures themselves poll well, and they have been getting more popular, not less. Resolve Strategic had support for the CGT changes at 36–21 in mid-May and 35–21 in its 9–15 August field; negative gearing at 35–21 and then 35–20. Support for the idea of reducing house prices ran at 60–10. Essential in late May had the property measures at 33% support to 27% opposition. The one component that polls badly is the discretionary trust change, at 26–38.

What polls catastrophically is not the policy. It is everything wrapped around it.

Freshwater Strategy found 54% saying the government had broken a promise. Newspoll put the budget’s net economic score at −25 and its personal impact at −41, with 52% expecting to be worse off against 11% better off. One published survey had voters rating it the worst budget since 1993 — the “L-A-W” budget, which is the canonical Australian broken-promise reference and not a coincidence. Albanese’s net approval has fallen from −17 in May to around −22 in August and −23 in early September.

That asymmetry is the central political fact of the whole debate, and it cuts against the repeal case rather than for it. An opposition can win an election on “they lied to you”. It cannot easily convert that into a mandate to restore a concession that a plurality of voters supports removing — particularly when its own attack line is generational. Tim Wilson’s framing is that the government protected “those who’ve accumulated wealth” while it “kneecap[ped] young Australians”. It is a good line. It also sits awkwardly beside a policy of handing investors their deductions back.

There is one more thing the polling does not contain, and its absence should be noted rather than filled in with assumption: we could not find a single published poll asking Australians directly whether the changes should be repealed. Anyone telling you a majority wants them gone is telling you something no pollster has measured.

The $2 a week problem

Treasury’s budget-night factsheet said the changes would produce “an expected increase of less than $2 per week for a household paying the current median rent”. It is the most quoted number of the whole reform and it has become the government’s biggest liability — not because it has been disproved, but because it can no longer be defended in public.

Start with what the number actually means, because almost every use of it since May has been wrong. Treasury was not forecasting total rent movement. It was estimating the marginal effect of the tax change, holding everything else constant — interest rates, migration, vacancy rates, construction costs. It is a difference between two modelled worlds, not a prediction about your lease.

Now look at what happened next. In the June quarter of 2026, the combined-capital median house rent rose $20 a week. Sydney rose $50 — its sharpest quarterly increase in four years — Brisbane $20, Adelaide and Canberra $10 each, Darwin $40. Every capital recorded a record median. National vacancy sat at 1.6% on Cotality’s measure and 1.3% on SQM’s, with every capital under 2%. Cotality had annual rent growth at 5.9% against wage growth of 3.3%, and rents consuming close to a third of household income — a record share.

Because the political point is this: a marginal estimate that cannot be observed, published against a backdrop of $20-a-week actual increases, is not a number that can win an argument. It is unfalsifiable in practice. Every rent rise between now and the election will be hung on it, and the correct rebuttal — “that is not what the modelling measured” — is a sentence no treasurer wants to have to say twice.

The professional estimates are also nowhere near each other, which tells you how little anybody really knows.

The two bottom bars come from a specific and testable claim. NAB’s Gareth Spence put it this way: for a Sydney or Melbourne investment property, lifting the gross rental yield a full percentage point — from about 3.5% to 4.5% — implies rents rising 25–30%. SQM’s Louis Christopher independently estimated that investors would need 1 to 1.5 percentage points of extra yield to be compensated, which is “about a 30% increase in rents”, or roughly 15% if the adjustment is shared with lower prices.

Peter Downes of Outlook Economics, a former Treasury economist, modelled it through the AUS-M macroeconometric model for the Senate inquiry and got rents about 6% above baseline by 2029 — roughly $2,000 a year, or $38 a week. His framing of the timing is the sharpest thing anyone has said about this policy in either direction.

Against that, the counter-argument is not weak. More than 80% of investor mortgages buy existing homes, so an investor exiting usually sells to somebody who then does not need to rent — the dwelling and the household leave the rental pool together. Victoria raised investor holding costs in 2024 and saw an investor exodus, yet its rent growth over five years came in below the national average. And the observable data so far does not support the sell-off story at all: new listings in the four weeks to 5 July 2026 were 6.2% below the five-year average. The market is softening because buyers have withdrawn, not because sellers have flooded in.

The ABS was careful about the order of causes when it published the lending figures: the cash rate first, the budget second. That ordering is doing a lot of work, and it is the right ordering. But an investor-loan fall of 8.6% in a single quarter, the largest since September 2022, is the one number in the whole dataset that looks like people responding to a tax rather than to a mortgage rate.

Who would actually protest

Assume the rent effect is real and larger than Treasury says. Who turns that into political force? The answer is less obvious than it looks, and for two of the three groups it is close to nobody.

Renters: almost certainly not — and not for the reason you would guess

About 30.6% of Australian households rented at the 2021 Census, and the growth is entirely in the private market: private rental went from 20% of households in 1999–2000 to 26.2% in 2019–20 while social housing halved from 6% to 3%. It is also no longer a life stage. The AIHW finds the rise in renting is most pronounced among households headed by people aged 35–49 and 50–64. A renter bloc concentrated among twenty-somethings can be dismissed as transitional. One thickening through the middle of the age distribution cannot.

So the constituency exists. What it has never had is a habit of mobilisation. The only period of mass renter direct action in Australian history is the anti-eviction movement of 1929–36, and its wins were state-level and administrative. Modern renter politics runs through submissions, media and Senate inquiries. There has been no rent strike and no protest of comparable scale in the current crisis.

But the decisive point is not capacity. It is that the organised renter movement has already told us how it will read rising rents, and it is not as an argument for repeal.

That statement came from a release titled Politicians must stop using renters to defend investor tax breaks, published with rents rising and the market falling. The Senate inquiry into the CGT discount, which reported on 17 March 2026, went further and explicitly warned against grandfathering and carve-outs on the ground that they would weaken the affordability impact. Everybody’s Home welcomed the passage of the bill in June and immediately called for more.

So rent inflation, in the hands of the organised renter sector, becomes evidence that the reform did not go far enough. It is worth noticing that this is an assertion, not a measurement — they have produced no more data than the industry has. But it means the political risk from rising rents does not run through renter organisations at all. It runs through unorganised renters as voters, and through the industry amplifying every rent print, which is a slower and much weaker transmission mechanism.

One more thing to discount: the idea that renters do not turn out. Australia has compulsory voting and 98.2% enrolment. The renter participation deficit is at the enrolment and roll-currency margin, driven by residential mobility, and it is worth a few points among the young and mobile. Anyone telling you renters do not vote is importing an American fact.

Homeowners: no, and there is no precedent at all

This is the strongest null finding in the whole exercise. Australian homeowners have never organised against falling house prices. Not in the 1890s, when Melbourne fell 51% in real terms. Not in 1974, when Sydney fell 18% and Melbourne 24% in real terms with no recovery until 1988. Not in the 1990–91 recession. And not in 2017–19, when parts of Sydney and Melbourne fell more than 20% immediately after an election fought partly on negative gearing — a fall which produced no movement and almost no negative equity, because only about 2¾% of securitised loans by value were underwater and roughly 90% of that was in the mining-bust states rather than the price-fall cities.

Homeowners express price falls through the ballot box and through what they spend, not through organisation. There is one live detail worth watching, though: in the current downturn the upper quartile is falling about three times faster than the bottom, and Sydney is worst. That is the geography of Liberal heartland and teal seats rather than the geography of renters. If a homeowner backlash ever does materialise, that is where it will be — and it will have no playbook to draw on.

Builders, developers and agents: yes — but the coalition has already split

These are the only organised, funded and experienced campaigners in this fight. The benchmark is the 2010 mining tax campaign: $22 million of industry advertising, and a prime minister gone. The property industry’s real weapon is different and arguably better. In 2016 more than twenty real estate chains coordinated a campaign against Labor’s negative gearing policy whose distinguishing feature was not spend but distribution — roughly 80,000 licensed agents and property managers with a direct, recurring, trusted relationship with essentially every tenant and every seller in the country. A property manager telling a tenant their rent is rising because of the CGT change is a medium the Minerals Council never had.

On 23 March 2026, seven weeks before the Budget, the Housing Industry Association, Master Builders, the Property Council and the Real Estate Institute issued a joint statement as a single bloc, with commissioned modelling putting the cost at up to 45,943 fewer dwelling starts by 2029.

What happened after the Budget is more interesting than the campaign before it. The coalition fractured along precisely the line the policy drew.

The naive prediction was that builders win and agents lose. What actually happened is that UDIA — the purest new-supply body — was the only one to welcome the measure, and everyone else converged on the HIA’s argument, which is the best single line anyone has produced in this debate:

That is an empirically testable claim and it is the crux. If a falling established market destroys buyer confidence, land values and pre-sale finance, the new-build exemption is worth less than it looks on paper. If it does not, the exemption does exactly what it was designed to do.

The most telling fact in the entire record, though, is what has not happened. There has been no paid advertising campaign against the 2026 measure on the 2010 or 2016 model. The Property Council is submitting on exposure drafts. That is the behaviour of a lobby that has decided the principle is lost and the money is now in the detail.

Would there be less supply?

Probably somewhat. But the measure is nowhere near the main thing happening to Australian housing supply, and the current data makes that unusually clear.

Dwelling commencements fell 11.2% in the March quarter of 2026 — two months before the Budget — with higher-density commencements down 20.7%. At the same time higher-density approvals were up 19.9% year on year. Approvals rising 20% while commencements fall 21% is not a tax-incentive problem. It is a finance and feasibility bottleneck between approval and shovel, and it predates the announcement.

The rest of the picture points the same way. Construction costs are about 51% above pre-pandemic and re-accelerating — house construction output prices rose 5.9% over the year to June 2026, the fastest since mid-2023, on fuel costs and bricklayer and carpenter shortages. Commencements are running at roughly 192,000 a year against about 240,000 needed. The National Housing Supply and Affordability Council pushed its Accord completion date from September 2030 to December 2030 in its August update, and named softer sentiment and interest rate rises as the deferral drivers, not tax. Two years into a five-year target, about 308,000 of 1.2 million homes are done: a quarter of the target in two-fifths of the time.

And one figure cuts squarely against the crisis framing: construction insolvencies fell in 2025–26, to 3,472 from a higher base, the first fall in five years, within a total company insolvency count that also declined.

Will it actually be inflationary for rents?

The best-identified Australian evidence says no, and the honest version of that answer includes why it might still be wrong.

In October 2024 the Reserve Bank published a study by Twohig, Yadav and Hambur using anonymised individual tax return data from 2006–07 to 2018–19, comparing rental income growth for investors in the same local area — which nets out the demand shock that moves everybody’s rents at once. The finding: for every extra dollar of mortgage interest cost, investors raise rents by about one cent, rising to three cents when rates are climbing. To put a number on it, the median investor’s interest costs rose roughly $850 a month between April 2022 and January 2024, implying about $10 a month of extra rent — under half a per cent. Their conclusion: the level of housing demand relative to the housing stock is what drives rents.

Apply that to this measure and the mechanism looks weaker still, because of the grandfathering. Quarantining raises after-tax holding costs only on newly acquired established property. Everything held on 12 May 2026 is untouched. The affected share of the rental stock therefore grows by a percentage point or two a year. A pass-through effect confined to two per cent of the stock cannot be highly inflationary for the other ninety-eight.

If there is a rent effect, it has to run through investor exit shrinking the rental stock rather than through pass-through. And that channel has a well-known problem: every dwelling an investor sells is bought by somebody. If the buyer is an owner-occupier, a rental property and a renting household leave the pool together. Treasury’s whole case rests on about 75,000 such buyers over a decade.

The odds, scenario by scenario

What follows are richer.au’s own estimates. They are not published forecasts, no bookmaker is pricing this, and we could not find a single economist or political scientist who has put a number on the repeal question in print. So treat these as a structured argument with the working shown, not as data. The horizon is the end of the parliament elected at the next election — roughly 2031.

What the history changed. Our first cut of this table put a repeal at about one chance in eleven and gave scenario B 31%. Reading the 1985 episode properly moved weight in one specific direction: not toward a Coalition repeal, but toward the government retreating from its own measure. That is what actually happened in 1987, without an election intervening, driven by Sydney rents and a state poll. Scenario B is up to 35% and now explicitly includes a deferral of the 1 July 2027 start date, which is the cheapest political fix available to a government facing a rental crunch in an election year. Scenario E is up to 7% because the organised renter sector has spent 2026 arguing the reform was too soft, not too hard. Gate one is nudged to 22% on the strength of the industry’s distribution network. The headline number moves only from 9% to about 10% — because the most likely form of retreat was never a repeal.

Three things fall out of that table that are worth more than the numbers themselves.

First, the most likely change is not a repeal — it is a softening by the people who wrote it. That process has already started. A small-business carve-out was conceded on 18 June, five weeks after budget night. In August the government agreed to fast-track a fix for the so-called widow tax, where a surviving co-owner who inherits their partner’s share is treated as making a new acquisition and loses the grandfathering — the same defect catches people separating, including after family violence. Taylor made supporting the government’s $37 billion NDIS savings package conditional on fixing the “cruel” loophole immediately; Chalmers said he was up for the discussion. That is an opposition negotiating improvements to a scheme it has promised to abolish, which is a fairly reliable tell about what it expects to happen.

There is also an available middle position that no major party currently occupies. The expert criticism of this package is remarkably consistent and remarkably narrow: economists like indexation and dislike the 30% minimum tax bolted onto it. Michael Brennan of the e61 Institute, a former Productivity Commission chair, called indexation “a principled approach” that ought to be at the heart of any capital gains system — and opposes the 30% minimum. Saul Eslake takes the same split.

Second, the grandfathering makes a repeal worth less to the people who would campaign for it — though, as the 1985 section above sets out, grandfathering alone is not what separates this attempt from the last one. Labor lost the 2019 election proposing negative gearing and CGT changes from opposition, and dumped the policy in 2021. The lesson it drew is visible in the architecture: legislate from government, protect everyone who already owns something, and let the affected group accumulate one purchase at a time. Compare that with the carbon tax, which was repealed inside a year. The carbon tax had an immediate, visible, universal cost with an organised and angry constituency. This measure’s losers are a slowly growing cohort of new investors who are not yet paying anything, and whose interests run directly against a much larger group of prospective first home buyers. That is a far worse base to run a repeal campaign from.

Third, the constituency is smaller than the noise suggests. The ATO’s latest taxation statistics show 2.26 million individuals with a rental property interest, of whom 1.12 million are negatively geared. But the measure only touches people who buy established residential property after 12 May 2026 — Treasury put that at roughly 1% of taxfilers a year, about 230,000 people. Meanwhile the RBA counts about 2.3 million housing investors in total, nearly 40% of them in the top income quintile, with a median age that has risen from 45 to 51 over two decades and more than a quarter now over 60. A repeal campaign is asking a young electorate to restore a concession that flows disproportionately to older, wealthier households. It can be won. It is not an easy win.

What a reversal would — and would not — undo

Suppose it happens. This is the part almost nobody has thought through, and it is where the practical money is, because a repeal does not restore the world of 11 May 2026. Several things are now permanent regardless of what any future parliament does.

The 1 July 2027 valuation point is probably permanent

If the indexation regime runs from 1 July 2027 and is repealed in, say, 2029, you do not get a clean slate. You get a cost base in three layers: growth to 30 June 2027 under the old 50% discount, growth from 1 July 2027 to the repeal under indexation plus the 30% minimum, and growth after the repeal under whatever replaces it. The apportionment you have to perform at 1 July 2027 — by valuation, or by the formula in the second tranche — still has to be performed, because it fixes the boundary between the first two layers. Repeal removes the future rule. It cannot remove the past one. Keep the valuation evidence for every CGT asset you hold across that date, whatever you think will happen politically.

A prospective repeal strands one cohort permanently

This is the single most important unanswered question in the debate, and it is genuinely unanswered: the Coalition has made no public statement on whether its repeal would restore negative gearing to people who bought between 12 May 2026 and the repeal date. Tim Wilson has said Australians “would still be able to negatively gear investment properties under current rules” under a Coalition government, which is ambiguous about which “current rules” he means. Nobody has said the word “retrospective” in either direction.

The arithmetic makes this unavoidable. By the earliest election date, the affected cohort will have been buying for roughly fifteen months. A prospective-only repeal leaves them permanently worse off than both the people who bought before them and the people who buy after them — which is politically awkward for a party campaigning on the injustice of the change. A retrospective repeal costs revenue and invites the charge of retrospective tax legislation, which both sides of politics have spent decades saying they oppose. There is no comfortable answer, which is presumably why there has not been one.

Prices and rents do not simply spring back

Announcement effects reverse faster than realised ones. The clearest measurable response to date — the 8.6% quarterly fall in investor loan commitments — is sentiment, and sentiment can turn on a speech. But if the measure has been running for a year or two before repeal, some of the adjustment will have been embedded in transactions: properties bought at lower prices, leases reset at higher rents, portfolios restructured into new builds or out of residential property altogether. Those are not undone by a bill. CBA’s estimate is that the tax changes subtract about 0.6 of a percentage point from price growth by end-2026 and just under a full point over 2027, for a long-run level effect of around 3%. A repeal in 2029 recovers part of that. It does not recover the transactions.

Someone has to pay for it

The $3.63 billion over the forward estimates is committed to the Working Australians Tax Offset. Beyond the forwards the number grows sharply as the grandfathered stock turns over — how sharply is genuinely unclear, and the published figures do not reconcile: $77.2 billion over a decade has been reported for the broader package, the budget papers show $3.63 billion to 2029–30 for this measure, and Grattan has estimated the policies improve the budget by more than $20 billion a year within a decade. Those three numbers are not measuring the same thing. What is not in doubt is the direction: the longer the measure runs, the more expensive repealing it becomes, and the harder it is to fund out of an opposition’s costings.

The compliance machinery is already built

The ATO has $90.7 million over five years to implement this, Treasury $8.1 million over two. The Office of Impact Analysis put average regulatory costs at $88.4 million a year. Taxpayers who obtain valuations at 1 July 2027 will have paid for them. None of that comes back.

What to do while it is unresolved

The practical conclusion is not “wait and see”, and it is certainly not “buy now because it will be reversed”. It is narrower than either.

Stamp duty decides whether you can buy. Land tax decides whether you can hold. And the tax treatment of the loss decides only how much of somebody else’s money is subsidising a decision you should have been able to justify without it. The people who stay Richer through a change of regime are the ones whose numbers worked before the deduction was applied — and who wrote down what the asset was worth on 1 July 2027 while everyone else was arguing about the election.

Frequently asked questions

Sources

All figures are as published to 13 September 2026. Polling figures are point estimates from individual polls, not averages, and different houses use different methods. Scenario probabilities are richer.au’s own estimates and are not published forecasts.

1. Federal Register of Legislation, Treasury Laws Amendment (Tax Reform No. 1) Act 2026, Act No. 49 of 2026, Royal Assent 26 June 2026.
2. Australian Taxation Office, “Tax reform — Boosting home ownership — Reforming negative gearing and capital gains tax”, new legislation guidance.
3. Australian Government, Budget 2026–27, factsheet “Negative Gearing and Capital Gains Tax Reform”, 12 May 2026.
4. Australian Government, Budget Paper No. 2, 2026–27, measure “Reforming negative gearing and capital gains tax” — receipts of $3,630m to 2029–30.
5. The Hon Dr Jim Chalmers MP, second reading speech, Treasury Laws Amendment (Tax Reform No. 1) Bill 2026, 28 May 2026.
6. ABC News, “Will CGT and negative gearing budget changes make housing cheaper for first home buyers?”, 13 May 2026.
7. ABC News, “Government breaks promise with restrictions to negative gearing and capital gains tax discount in federal budget”, 12 May 2026.
8. ABC News, “Coalition to repeal Labor property tax changes”, 15 May 2026 — budget reply of 14 May 2026.
9. Commonwealth of Australia Constitution s13 and s28, and the Commonwealth Electoral Act 1918 — election timing for the 48th Parliament.
10. Newspoll, field 14–17 May 2026 (n=1,252) and 4–8 August 2026 (n=1,242).
11. Freshwater Strategy, field 13–15 May 2026 (n=1,384).
12. Resolve Strategic, field 13–16 May 2026 (n=1,800) and 9–15 August 2026.
13. DemosAU / Capital Brief, field 16–18 June 2026 (n=1,497).
14. YouGov Public Data poll, field 7–14 July 2026 and 1–8 September 2026; Roy Morgan, field 31 August – 6 September 2026.
15. Essential Research, field late May 2026 (n=1,027).
16. Domain Rental Report, June quarter 2026 — combined capitals house median +$20/week; Sydney +$50, +6.3%.
17. Cotality Quarterly Rental Review, June quarter 2026 — national median $705/week, +5.9% annual, vacancy 1.6%; SQM Research vacancy series, July 2026.
18. Outlook Economics (Peter Downes), submission to the Senate Economics Legislation Committee inquiry into the Treasury Laws Amendment (Tax Reform No. 1) Bill 2026, June 2026 — AUS-M modelling, rents ~6% above baseline by 2029.
19. NAB (Gareth Spence) and SQM Research (Louis Christopher), reported August 2026 — yield-recovery estimates of 25–30%.
20. Cotality Home Value Index, August 2026 — national −0.9% for the month, −3.6% from the March 2026 peak, 93% of capital city suburbs falling.
21. Australian Bureau of Statistics, Lending Indicators, June quarter 2026 — investor commitments −8.6%, first home buyers −2.9%.
22. Australian Bureau of Statistics, Building Approvals, July 2026 — private dwellings excluding houses +19.9% annual.
23. Commonwealth Bank of Australia, “2026 Budget: updated housing outlook”, 18 May 2026 — long-run price level effect ~3%.
24. Grattan Institute, “The budget we’ve been waiting for”, 13 May 2026.
25. ABC News, “Economists defend capital gains changes”, 15 June 2026 — Michael Brennan, Saul Eslake, Peter Varela, Paul Keating.
26. ABC News, “Labor to fast-track widow tax fix”, 18 August 2026.
27. ABC News, “Greens back CGT and negative gearing changes”, 23 June 2026.
28. Australian Taxation Office, Taxation Statistics 2022–23 — 2,261,080 individuals with a rental interest, 1,117,175 negatively geared.
29. Reserve Bank of Australia, Bulletin, “Insights from New Data on Australian Housing Investors”, May 2026.
30. Australian Electoral Commission, Farrer by-election, 9 May 2026 — One Nation 57.55% two-candidate-preferred.
31. Office of Impact Analysis, “Capital Gains Tax and Negative Gearing”, published 21 May 2026 — average regulatory costs $88.4m per year.
32. KPMG, “Consultation on second tranche of capital gains tax and negative gearing reforms legislation”, 6 August 2026.
33. Philip Soos, Written Off: Negative Gearing, Prosper Australia, October 2012 — ABS CPI rents by capital, CPI-deflated, 17 July 1985 to 1 July 1987.
34. O’Donnell, “Quarantining Interest Deductions for Negatively Geared Rental Property Investments”, eJournal of Tax Research, 2005 — mechanism, grandfathering and the 1987 reversal.
35. Australian Taxation Office, Consumer Price Index series — CPI rose 18.6% between June 1985 and June 1987.
36. Australian Bureau of Statistics, Cat. 8731.0, “Changes in the mix of dwelling types”, 5 July 2002 — dwelling completions 1984–85 to 1989–90.
37. Bob Hawke, address to the AFR post-Budget dinner, 16 September 1987 (PM Transcripts 7223).
38. ABC Fact Check, “Did abolishing negative gearing push up rents?”, 6 May 2015 — Keating’s 1987 cabinet submission; Saul Eslake.
39. Grattan Institute, Hot Property: negative gearing and capital gains tax reform, April 2016.
40. Twohig, Yadav and Hambur, “Do Housing Investors Pass-through Changes in Their Interest Costs to Rents?”, RBA Bulletin, October 2024 — pass-through of about 1 cent in the dollar, up to 3 cents when rates rise.
41. Cho, Li and Uren, “Investment Housing Tax Concessions and Welfare: A Quantitative Study for Australia”, International Economic Review, 2024 — supply −1.8%, rents +2.5%.
42. ABS Census 2021, Housing; ABS Housing Occupancy and Costs 2019–20; AIHW, Home ownership and housing tenure, 16 October 2025.
43. Everybody’s Home, “Politicians must stop using renters to defend investor tax breaks”, 20 August 2026; Senate inquiry into the CGT discount, report 17 March 2026.
44. HIA, Master Builders, Property Council and REIA joint statement, 23 March 2026, with Qaive and Tulipwood Economics modelling; updated joint statement and modelling, 11 September 2026.
45. UDIA, Master Builders and HIA post-Budget statements, 12 May 2026; Property Council submission on the second-tranche exposure drafts, 28 August 2026.
46. ABS Building Activity, March quarter 2026; ABS Producer Price Indexes, June quarter 2026; NHSAC, State of the Housing System 2026 (30 April 2026) and Quarterly Report, August 2026.
47. Reserve Bank of Australia, Financial Stability Review, April 2019, Box B — negative equity in the 2017–19 downturn.
48. Reserve Bank of Australia: Simon, “Three Australian Asset-price Bubbles”, 2003; RDP 8904, “Removal of Interest Rate Controls”, 1989.
49. ASIC insolvency statistics, 2025–26 — 3,472 construction insolvencies, the first annual fall in five years.

Keep reading