Industry modelling says 10,700 fewer homes. Treasury says 30,000 more.

The Weekly Read · 15 September 2026

On 11 September 2026, four construction and property bodies released modelling of the government's housing tax package. It puts new home starts 10,700 lower over the next four years. Treasury's number for the same budget package runs the other way: a net 30,000 more homes over ten years. The rent forecasts are five times apart, about $10 a week against about $2.

Why does the same law produce opposite numbers? The answer first: on whether the tax changes reduce housing supply, the two sides are close. What splits them is how many homes they expect the government's $2 billion fund to deliver.

1. The number four industry bodies published

The statement came from Master Builders Australia, the Housing Industry Association, the Property Council of Australia and the Real Estate Institute of Australia. Qaive and Tulipwood Economics did the modelling for them.

Across the four financial years from 2026-27 to 2029-30, the cumulative result is this.

• New home starts down 10,700

• Rents up about $10 a week (by 2029-30, with no regional split)

• More than 4,700 construction jobs gone

• Gross domestic product about $1.05 billion lower

Four policies went into the model: the change to the rules that let rental losses be offset against other income, the change to the capital gains tax (CGT) discount, the ban on new borrowing by self-managed super funds (SMSFs) for residential property, and the government's $2 billion Housing Support Program, which is there to add supply. The 10,700 is the net of all four, and the breakdown by policy was not published.

2. The number Treasury published

Treasury costed the same budget package, but it did not measure the same list. Its estimate covers the rental loss offset change, the capital gains tax change and the budget's housing supply investment. The SMSF borrowing ban is not in it. That measure was added after the budget, through an agreement between Labor and the Greens.

Treasury's result is a net 30,000 additional homes over ten years: the tax changes take out 35,000, the supply investment adds 65,000. On rents it has a rise of about $2 a week for a household currently paying the median rent. Those figures appeared in the federal budget papers of 12 May 2026.

	Four industry bodies	Treasury
Period	2026-27 to 2029-30 (four years)	next ten years
In the model	rental loss offset change · capital gains tax change · SMSF residential borrowing ban · $2 billion fund	rental loss offset change · capital gains tax change · housing supply investment (SMSF borrowing ban excluded)
Housing supply	10,700 fewer starts	net 30,000 more (tax −35,000 + supply investment +65,000)
Rents	up about $10 a week	up about $2 a week

The periods differ and so does the scope. But the reason the results point opposite ways is somewhere else.

3. Why the numbers run opposite: how much you trust the fund

A statement the same bodies issued on 29 May 2026 shows how far apart the two sides sit on one figure: how many homes the $2 billion fund delivers across four years. The industry number is 5,300. Treasury's is 26,000.

Take the industry calculation and swap that one figure for Treasury's. The fall of 10,700 becomes a rise of about 10,000.

Over four years	Industry calculation	With Treasury's fund figure
Tax changes + SMSF borrowing ban	−16,000	−16,000
$2 billion fund	+5,300	+26,000
Total	−10,700	+10,000

The −16,000 is not a figure the industry published. We worked it out by taking the fund's 5,300 out of the headline −10,700. Two assumptions sit inside that: that the September modelling used the same fund figure as May, and that the policy effects can simply be added together. The September material does not give a fund figure of its own.

On the tax side the two are close. The industry has about 16,000 fewer homes across four years, around 4,000 a year. Treasury has 35,000 across ten, around 3,500 a year. The periods and the scope differ, so it is not a clean comparison, but the scale is much the same.

So what the two sides are actually arguing about is not how many homes the tax changes cost.

It is how many homes government money buys.

Why do they see that so differently? Divide the $2 billion by the homes each side expects.

• Treasury: $2 billion ÷ 26,000 homes = about $77,000 a home

• Industry: $2 billion ÷ 5,300 homes = about $377,000 a home

You cannot build a house for $77,000. Treasury is not treating the money as something that builds houses directly. It is treating it as the roads and the water and sewer connections that were missing, with private developers building the homes once those are in. The industry expects about a fifth as much development to follow. The May industry statement calling the money a Local Infrastructure Fund fits that reading. Neither side published its method, so this explanation is ours, drawn from the published numbers.

4. Why the rent forecasts sit five times apart

The industry argument runs like this. Tax benefits shrink, so investors sell. An owner-occupier buys the property, so one rental home is gone. Rents rise.

The problem is that nobody has properly measured how often that chain actually completes.

Take the survey the industry cites, run by the Property Investment Professionals of Australia (PIPA). The share of investors who had sold in the previous 12 months rose from 12.1% in 2023 to 16.7% in 2025. In the same survey, the share calling the next 12 months a good time to invest also rose, from 45.7% in 2024 to 59% in 2025. And the properties that sold went to other investors (42%) more often than to owner-occupiers (37%), on a question that allowed more than one answer. Where another investor buys, the home stays in the rental pool. The survey also draws on people who chose to respond, so it cannot be read as a picture of investors as a whole.

The Reserve Bank did analyse investor data in May 2026, but what that work covers is investor age, leverage and location. Whether the homes investors sell actually leave the rental market is not in it.

What this changes for buyers

1. Do not decide on a headline number. The 10,700 and the 30,000 are both forecasts of what a fund that has not been spent yet will deliver. Which one was closer will not be visible until around 2030.

2. Look at the rules already settled instead.

Rule	When it applies	Who needs to look again, and at what
Ban on SMSF borrowing for residential property	Commenced 10 August 2026. A contract signed before that date is protected even if settlement or loan approval falls after it, and refinancing is protected within the existing limit	If a fund was planning to buy a home, the structure needs a second look. Borrowing for business real property, a shop or a warehouse, is still available
Restriction on offsetting rental losses	Commences 1 July 2027. Applies to property acquired after 7.30 pm on 12 May 2026	An investment property bought now is inside it. Any cash flow worked out on the assumption of a tax refund needs redoing
Capital gains tax change	Applies to gains arising after 1 July 2027. Gains arising before then keep the existing 50% discount	New builds and affordable housing can choose between the existing discount, up to 60% for affordable housing, and the new method. At the same price, a new build and an established home no longer return the same after tax

The four years the industry modelled start on 1 July 2026, and for the first of them the only rule actually in force is the SMSF borrowing ban.

3. What East Point takes from it

• Investors: whether a property stands up on rent alone, with no tax benefit in the calculation, now matters far more than it did. Before buying, work the after-tax return through the new rules with your accountant or tax agent.

• Owner-occupiers: the hope that investors stepping back means less competition is not well supported yet. Selling investors and buying investors are rising at the same time.

4. What we will be watching

How much of this fund is allocated to South East Queensland, and whether building approvals and starts actually rise behind it. We check that across the eight South East Queensland councils every month. Those numbers will tell us which forecast was closer long before the forecasts themselves are settled.

---

Sources

• Joint industry modelling (10,700 fewer starts, rents up about $10 a week, more than 4,700 construction jobs, gross domestic product about $1.05 billion, and the list of policies modelled): joint statement by Master Builders Australia, the Housing Industry Association, the Property Council of Australia and the Real Estate Institute of Australia, 11 September 2026. Modelling by Qaive and Tulipwood Economics

• Treasury estimate (net 30,000 more homes, made up of 35,000 fewer from the tax measures and 65,000 more from the supply investment; rents up about $2 a week): federal budget papers, 12 May 2026. Cross-checked against ABC News reporting of 13 May 2026

• Fund assumptions compared (5,300 against 26,000, both over four years): joint statement by Master Builders Australia, the Property Council of Australia and the Real Estate Institute of Australia, "Post Budget Modelling Highlights Impact on Housing Supply and Rents", 29 May 2026. That statement refers to the fund as the Local Infrastructure Fund

• The −16,000 for tax and SMSF measures, the per-home figures ($77,000 and $377,000), and the +10,000 net effect of swapping the fund assumption: East Point calculations from the figures above

• Ban on new SMSF limited recourse borrowing arrangements for residential property, with the commencement date, the protection for existing contracts and the business real property carve-out: Treasury Laws Amendment (Tax Reform No. 1) Act 2026 (assented 26 June 2026, commenced 10 August 2026). Conditions as reported across SMSF trade publications

• Commencement dates for the rental loss offset restriction and the capital gains tax change, and the new build and affordable housing exception: Australian Taxation Office new measures guidance, and ABC News reporting of 12 May 2026

• Investor selling, where sold properties went, sample recruitment, and the "good time to invest" share: Property Investment Professionals of Australia (PIPA) Annual Investor Sentiment Survey

• Scope of the housing investor data: Reserve Bank of Australia Bulletin, May 2026, "Insights From New Data on Australian Housing Investors"

---

General information only, not tax, legal or financial advice. Nothing here is financial product advice or a recommendation about any purchase or buying structure. Confirm your own position with your accountant, tax agent or licensed financial adviser before acting.