This submission addresses two of the instruments under consultation. On the draft Income Tax Assessment (Method for Apportioning Capital Gains and Capital Losses) Determination 2026, we identify two defects in the prescribed method as it applies to real property — the treatment of capital expenditure incurred after 1 July 2027, and of discontinuous value events such as rezoning — each of which converts genuine post-reform gains into concessionally taxed pre-reform gains. On the draft determination defining a “new residential dwelling”, our conclusion, for the reasons set out in section 2, is that the concession cannot be cured by amendment.
We make four recommendations:
- Recommendation 1.1 — The new residential dwelling concession should be removed in full. The loss quarantining rule and the new CGT treatment should apply uniformly to all residential investment property, regardless of when the dwelling was built.
- Recommendation 2.1 — Amend step 2 of the prescribed method so that the capital proceeds are first reduced by cost base expenditure (second to fifth elements) incurred on or after 1 July 2027, before the total growth rate is computed.
- Recommendation 2.2 — Following the UK approach, exclude from the method real property that, between 1 July 2027 and the realisation event, has been the subject of rezoning or comparable planning uplift. Such assets should determine their 30 June 2027 value by valuation.
- Recommendation 2.3 — The Determination should anchor the 30 June 2027 value of land to the applicable statutory valuation, as the default or a rebuttable presumption.
Recommendation 1. The ‘new residential dwelling’ concession should be abandoned
1.1 Summary
Prosper Australia’s position is that the new residential dwelling concession is bad policy and should be scrapped in its entirety. The scheme will prove incredibly complicated, international case studies of similar programs have a terrible track record, will hamper choice, will subsidise vacancies, dereliction and the destruction of homes in the affordable segment of the housing market.
Recommendation 1.1 — The new residential dwelling concession should be removed in full. The loss quarantining rule and the new CGT treatment should apply uniformly to all residential investment property, regardless of when the dwelling was built.
1.2 A design with a poor international record
The concession belongs to a well-studied family: tax preferences for investors who purchase newly constructed rental dwellings. The family’s record is one of high cost, weak additionality and eventual abandonment.
France is the fullest experiment, having cycled through the Robien (2003), Borloo, Scellier, Duflot and Pinel (2014) schemes. The Robien era left provincial towns with vacant, tax-driven investor stock in places without tenant demand. Pinel, the most refined version — with price ceilings, rent caps and geographic zoning the present draft does not even attempt — was evaluated jointly by the Inspection générale des finances and the CGEDD in 2019, which found weak additionality, a high fiscal cost per dwelling relative to alternatives including direct social provision, and substantial capture of the benefit in prices and developer margins.1 The Cour des comptes repeatedly criticised the cost and undemonstrated effectiveness of the rental investment tax expenditures.2 France declined to renew Pinel beyond 31 December 2024. It bears emphasis that the failed French schemes were more disciplined than this proposal: the draft instrument contains no value cap, no rent condition and no locational criterion, so a $6 million harbourside penthouse with a fresh certificate of occupancy qualifies on identical terms to an outer-suburban townhouse.
New Zealand supplies the closest structural analogue. Its 2021 interest limitation rules — the New Zealand version of loss quarantining — exempted “new builds” for 20 years, with eligibility keyed, as here, to the date a code compliance certificate issued. The carve-out required successive rounds of definitional guidance and boundary rulings, and the entire regime, carve-out included, was repealed within three years, with full interest deductibility restored from 1 April 2025.3
The supply of dwellings is constrained by land, planning and absorption rates, so a tax preference for purchasers of new dwellings raises willingness to pay before it raises output. The value of the preference is competed into the prices of new stock and, through the residual, into land values — accruing to landowners and developers rather than to renters or to net supply. It also re-arms leveraged investors against first home buyers in precisely the market segment — new dwellings — into which first home buyer programs steer them.
The concession also bears directly against the affordable end of the rental market. The affordable segment consists significantly of depreciated dwellings — yesterday’s mid-market houses and flats drifting down the rent distribution as they age — and filtering of this kind is the principal means by which private markets supply low-income housing.4 The package taxes that mechanism at both ends. Investment in the established, depreciated stock that actually houses low-income renters bears loss quarantining and the new CGT treatment in full, while the exemption is reserved for new dwellings, which enter at the top of the local rent distribution and reach the bottom only after decades of depreciation — a descent the instrument itself retards, since the personal, non-transferable status examined in section 1.4 penalises the turnover through which new stock ages into cheap stock. At the bottom of the market, where building value is smallest relative to land, the rational responses to that differential are precisely the behaviours the instrument elsewhere rewards: under-maintenance, dereliction and demolition (section 1.3), each of which extinguishes a cheap dwelling today in exchange for a concession-bearing dwelling that will not be cheap for a generation. A measure justified as an addition to supply operates, at the affordable margin, as a subsidy to finance the removal of the dwellings the poorest renters occupy and replacing them with dwellings those renters cannot afford.
1.3 A subsidy for dereliction and vacancy
The basic case in subsection 5(2) requires that the entity acquired an interest in land that “does not have a residential dwelling constructed or installed upon the land” at the acquisition time. The EM extends this to land occupied by condemned or derelict buildings, on the basis that premises unfit for habitation are not a “residential dwelling”. By contrast, an entity that acquires a sound but ageing house, demolishes it and builds a replacement obtains nothing: the basic case fails at paragraph 5(2)(a) because a residential dwelling existed at acquisition; the first special case fails because one-for-one replacement does not satisfy the “exceeds” test in paragraph 5(4)(c); and the third special case fails at paragraph 5(8)(a) because the building comprised a residential dwelling.
The consequence is that the habitability of an existing house at the instant of settlement determines whether the eventual replacement dwelling carries a lifetime tax concession. A vendor therefore maximises their sale price by ensuring the dwelling is uninhabitable at settlement, because uninhabitability is what unlocks the purchaser’s concession, and tax-motivated purchasers will pay a premium for dereliction. The draft rewards the deterioration of housing and penalises its maintenance. Nor is the boundary administrable. Neither the instrument nor the Act defines habitability; the test is whether the premises remain a “dwelling” within section 118-115, which the EM glosses as fitness “for human habitation” by reference to physical characteristics. That question would be self-assessed by the taxpayer, on facts fixed at a single moment and evidenced — if at all — by photographs and building reports, with a permanent tax status riding on the answer. Disputes are guaranteed.
The instrument also supplies a do-it-yourself version of the same pathway. Because paragraph 5(4)(c) compares the current number of dwellings on the land against the number at the acquisition time, an owner can manufacture the derelict outcome without a sale. An investor who buys a habitable house in 2028 (one dwelling at acquisition), allows it to deteriorate, demolishes it and builds two townhouses satisfies the first special case for both: two dwellings exceed one, and both certificates of occupancy post-date the acquisition and announcement times, so the exclusion in subsection 5(5) is not engaged. A neighbouring owner who maintains their house and builds one townhouse in the backyard adds the same single net dwelling — and receives one concession rather than two. Neglect outperforms maintenance.
Nothing in the basic case requires construction within any period of acquisition. Land acquired in 2028 and left idle until 2045 yields, on eventual construction, exactly the concession available to an owner who built immediately. For a measure whose statutory justification under subsection 26-160(4A) is “genuinely adding to… supply”, indifference between building now and building in two decades is remarkable: the concession is fully compatible with land banking.
The vacancy incentive recurs elsewhere in the draft. The transitional rule in paragraph 8(1)(c) preserves a developer’s deemed status only if the dwelling is held “at all times after the announcement time for the purposes of sale.” A developer who leases completed unsold stock — entirely ordinary commercial practice — presumably risks extinguishing the status; the tax-safe course is to leave completed dwellings empty.
1.4 The concession attaches to the wrong unit
The policy’s unit of account is the net addition to housing supply. The instrument’s unit of account is any dwelling bearing a certificate of occupancy first issued after the acquisition and announcement times.
Consider the EM’s own illustration of the first special case: an entity acquires land carrying an apartment block, demolishes it, and constructs a block containing more dwellings than before, whereupon “all the residential dwellings would be new residential dwellings for the entity.” Replace 100 units with 101 and the net addition to supply is one dwelling; the concessions granted number 101 — each a lifetime exemption from quarantining plus a CGT election. Meanwhile a one-for-one knockdown rebuild attracts nothing. Three consequences follow. First, the fiscal cost per marginal net dwelling is a large multiple of the headline concession. Second, the discontinuity invites token-unit gaming: appending a single studio to a redevelopment flips the entire project into concession status. Third, the treatment of economically similar projects becomes arbitrary.
The static drafting of paragraph 5(4)(c) creates a further leak. Because the test compares the number of dwellings on the land at any later time against the number at the acquisition time, once a site has ever exceeded its acquisition count by one, every subsequent replacement dwelling also qualifies. An owner acquires house A in 2028 (count: one) and builds townhouse B in 2029 (two exceeds one; B is new). In 2036 the owner demolishes A and builds C. At that time C is one of two dwellings on land that carried one at acquisition, and its certificate of occupancy post-dates both relevant times: C is a new residential dwelling, although it merely replaced existing stock.
1.5 Carve-outs breed carve-outs
Once “new” is carved out of the tax base, “new” must be defined; and each definition then requires its own exceptions, which require their own definitions.
A single activity — bringing a dwelling into existence — is split across four cases with materially different requirements. The independence limb (an interest “capable of being acquired… independently of any other dwelling”) appears in paragraphs 5(4)(d) and 5(8)(d) but not in 5(2)(d) or 5(6)(d). A duplex built on a single unsubdivided title therefore qualifies, or at least arguably qualifies, through the basic case, while the same physical outcome fails through the first special case; the EM’s stated policy on non-separately-titled dwellings (its granny flat discussion) is implemented in only two of the four cases. Mixed-use conversions cannot use subsection 5(8) at all — a building that “comprises” even a single caretaker’s flat is excluded by paragraph 5(8)(a) — but route instead through subsection 5(4), which, it should be noted, contains no construction or conversion requirement whatsoever. Identical works are thus tested under different requirement sets depending on whether the starting dwelling count was zero or one.
The anti-avoidance rule in subsection 5(10) then fails in both directions at once. The EM describes a provision confined to schemes that manufacture new-dwelling status “without having genuinely added to housing supply.” Those words are not in the provision. The operative text disregards a scheme wherever “a purpose of the entity, or one of the entities” — anyone in the chain — “is to treat a residential dwelling as a new residential dwelling to obtain a tax benefit”, with “scheme” bearing its usual, effectively unlimited, statutory meaning and “tax benefit” undefined in the instrument. Read literally, that describes every commercially rational participant in the very market the concession exists to create: developers will market these dwellings on their tax status, and purchasers will buy them for it. Because the provision is self-executing, every taxpayer must self-assess against it without the procedural protections of Part IVA, and an arm’s-length purchaser’s status can be destroyed by a purpose held by someone else entirely.
1.6 The rulebook this concession would require
A personal, non-transferable, potentially multi-decade tax status attached to individual dwellings — indeed to individual interests in dwellings, since the EM requires each interest an entity holds to be tested separately — cannot operate on the draft’s five sections. Consider what would have to follow.
- It would need preservation and rollover rules: for deceased estates, for relationship breakdown, for changes of trustee, for bare trustees and custodians (including limited recourse borrowing structures), and for entity restructures. The instrument contains none, so each of these ordinary events strips the status from the same physical dwelling mid-stream.
- It would need land rules: the EM asserts that “the land” in subsection 5(4) means the originally acquired parcel through any later subdivision, but the text does not say so, and consolidation, boundary adjustments and structures straddling former boundaries are unaddressed.
- It would need a verification regime for the second special case, because a purchaser’s eligibility under paragraph 5(6)(b) depends on the vendor’s status, which depends in turn on the vendor’s own acquisition history — facts the purchaser cannot observe or verify. That implies vendor certification at conveyancing, disclosure obligations and penalties for false statements: machinery of the kind the GST withholding and foreign resident capital gains withholding regimes required, none of which is provided.
- It would need certificate of occupancy rules: the definition (“however described”) spans eight jurisdictions’ inconsistent instruments — occupation certificates, certificates of classification, interim and partial certificates in staged developments — and everything turns on when a certificate was first issued, a fact a taxpayer may need to prove to the Commissioner from council records forty years after the event.
- It would need apportionment rules for part-year use under section 6 and for dwellings held through multiple interests with different statuses. It would need a rewritten subsection 5(10). And it would need ATO rulings, guidance and compliance products across all of the above, maintained indefinitely, for a status the Commissioner has no independent means of observing.
- And probably a lot more.
It is not worth going down this path. The concession will breed an extremely complicated tax structure that will prove the policy’s undoing, and may very well reduce the supply of housing, especially affordable housing, compared with simply quarantining losses without any concessions.
2. Apportionment
The draft Determination allows taxpayers to estimate an asset’s 30 June 2027 value by assuming it grew at a single compounding daily rate between purchase and eventual sale (s 5(3), steps 2–5). For the asset class the instrument names first — real property (s 5(2)(a)) — this assumption fails in the two situations that dominate real property in practice: where capital is spent improving the asset after 1 July 2027, and where value moves discontinuously, most importantly on rezoning. In both, the method certifies a 30 June 2027 value materially above the asset’s true value, converting genuine post-reform gains into deferred pre-reform gains that receive the 50 per cent discount and fall outside the 30 per cent minimum tax (deferred gains being exempt from Division 119).
Because the choice between the method and a market valuation is made at lodgement for the realisation year — with full hindsight — these errors are not symmetric noise. Taxpayers whose facts favour the formula will use it; all others will obtain a retrospective valuation. Only the revenue-negative errors survive. Two worked examples follow, each with the amendment that cures it.5
Defect 1 — Post-1 July 2027 expenditure contaminates the growth rate
Step 2 divides the capital proceeds from the realisation event by the first element of the pre-start date cost base. Where the taxpayer has incurred capital expenditure after 1 July 2027 — most commonly a renovation — the proceeds embed the value of that expenditure, but the divisor and the interpolation do not. The method therefore treats the improvement’s value as organic growth accruing smoothly across the entire ownership period, including years before the improvement existed. The same expenditure is then counted a second time, correctly, in the post-start date cost base at step 8, where it is indexed from the quarter incurred. One outlay thus appears twice in the taxpayer’s favour: once as phantom value inside the deemed 30 June 2027 figure (indexed from July 2027 — years before the money was spent), and once as indexed cost in the fourth element.
The enabling provision requires the determined method to “take into account … any expenditure including indexation in an element of the cost base” (s 112-185). A method that produces an identical deemed 30 June 2027 value whether a $700,000 renovation was undertaken in 2028, in 2033, or never (with a correspondingly lower sale price) takes no account of that expenditure in forming the one figure the method exists to produce.
Worked example — Substantial renovations to an investment property
Dana buys an established investment dwelling on 1 July 2019 for $500,000. Its true market value at 30 June 2027 is $800,000. In 2033 she undertakes a substantial renovation — a second storey and extension — for $700,000. (A substantial renovation of an existing dwelling does not add to supply and is expressly outside the “new residential dwelling” concessions; those rules neither apply nor interfere.) She sells on 30 June 2035 for $2,000,000, the renovation having returned only its cost. Half the days of ownership fall before 1 July 2027.
Her true economic position: a $300,000 gain accrued before the reforms; a $500,000 gain ($800,000 → $2,000,000, net of the $700,000 spent) accrued after — the gain the new regime exists to tax.
The formula: total growth rate = $2,000,000 ÷ $500,000 = 4.0; deemed 30 June 2027 value = $500,000 × 4.0^0.5 = $1,000,000, against a true value of $800,000. The $200,000 difference is the 2033 renovation, relocated by interpolation to before 2027.
Table 1: Renovated investment property — market valuation vs draft formula
| Market valuation (true) | Draft formula | |
|---|---|---|
| Deemed 30 June 2027 value | $800,000 | $1,000,000 |
| Deferred gain — 50% discount, exempt from 30% minimum tax | $300,000 | $500,000 |
| Post-start cost base (deemed value indexed 8 yrs + renovation indexed 2 yrs) | $1,710,200 | $1,953,800 |
| Post-1 July 2027 gain (new regime) | $289,800 | $46,200 |
| Tax on deferred gain (47%, after discount) | $70,500 | $117,500 |
| Tax on post-start gain (47%) | $136,200 | $21,700 |
| Total tax | $206,700 | $139,200 |
Dana runs both computations at lodgement — as the legislation permits, and as the ATO’s promised calculation tools will make trivial — and lodges the formula. She pays $67,500 less tax. Of her $500,000 in genuine post-reform profit, the new regime is shown $46,200: 91 per cent of the gain the reform was legislated to capture is invisible to it, re-emerging instead as discounted, minimum-tax-exempt pre-2027 gain. This is not avoidance; it is the instrument operating as drafted. The larger the renovation relative to the purchase price, and the longer the pre-2027 tenure, the larger the transfer.
Recommendation 2.1 Amend step 2 so that the capital proceeds are first reduced by cost base expenditure (second to fifth elements) incurred on or after 1 July 2027, before the total growth rate is computed.
Defect 2 — Non-uniform growth: the rezoning case
The method assumes value accrues at a constant compound rate. This might be a pragmatic solution in some cases. But for real property values it is not. This is particularly egregious because real property moves discontinuously on planning events. Where a rezoning or comparable decision confers uplift after 1 July 2027, the formula redistributes that uplift geometrically across the entire ownership period — attributing much of the dateable, post-reform, government-conferred windfall to the concessionally taxed pre-reform era.
When the United Kingdom introduced CGT in 1965 and faced the identical transition problem, time-apportionment was offered as the general shortcut — but land with development value was expressly excluded, on the stated ground that “the growth in value of the land over the period of ownership is unlikely to have been uniform,” with mandatory market-value rebasing at 6 April 1965 instead.6 The draft Determination does the opposite twice over: it names real property as the first eligible class, and it adds an after-the-fact election the UK never permitted.
Worked example — Farm rezoned
Ray buys two hectares of fringe rural-residential land on 1 July 2019 for $500,000. On 30 June 2027 it remains a paddock worth $600,000 — a figure appearing, at or about that date, on the relevant state valuation roll. In 2034 the land is brought within a precinct structure plan and rezoned for medium-density housing. On 30 June 2035 Ray sells to a developer for $4,500,000, never having spent a dollar on the land.
The formula: total growth rate = $4,500,000 ÷ $500,000 = 9.0; half the days of ownership fall pre-start, so the deemed 30 June 2027 value = $500,000 × 9.0^0.5 = $1,500,000 — two and a half times the paddock’s true value, because a 2034 planning decision has been smeared back to 2019.
Table 2: Rezoned farmland actual vs Treasury formula
| Market valuation (true) | Draft formula | |
|---|---|---|
| Deemed 30 June 2027 value | $600,000 | $1,500,000 |
| Deferred gain — 50% discount, exempt from 30% minimum tax | $100,000 | $1,000,000 |
| Post-start cost base (deemed value indexed 8 yrs) | $731,000 | $1,827,600 |
| Post-1 July 2027 gain (new regime) | $3,769,000 | $2,672,400 |
| Tax on deferred gain (47%, after discount) | $23,500 | $235,000 |
| Tax on post-start gain (47%) | $1,771,400 | $1,256,000 |
| Total tax | $1,794,900 | $1,491,000 |
Ray lodges the formula and pays $303,900 less tax. His concessionally taxed pre-reform gain is multiplied tenfold — $900,000 of a post-reform planning windfall reclassified as pre-2027 gain, discounted and exempt from the minimum tax. Every rezoning, structure plan and transport-corridor announcement between commencement and the eventual sale of affected land will reproduce this fact pattern; back-loaded windfalls always favour the formula, so affected owners will always elect it. A regime whose stated object is the fairer taxation of asset income will, in its first transitional act, deliver its most concessional treatment to the least earned gains in the property market.
Recommendation 2.2 Following the UK approach, exclude from the method real property that, between 1 July 2027 and the realisation event, has been the subject of rezoning or comparable planning uplift. Such assets should determine their 30 June 2027 value by valuation.
Recommendation 2.3 Unlike the United Kingdom in 1965, Australia already maintains comprehensive statutory land-valuation systems across the states and territories. For effectively every parcel of rateable land, there will therefore be an independently determined statutory land value in force, or a statutory valuation benchmark applicable, around the deemed-disposal date. These values are produced for state and local rating and taxation purposes rather than by taxpayers seeking to minimise a subsequent federal tax liability. The Determination should use the applicable statutory land value as the default anchor for the land component of the 30 June 2027 valuation, subject to prescribed adjustments where the statutory valuation date differs materially from 30 June and to a rebuttable presumption where the taxpayer can demonstrate that the statutory value materially departs from market value.
This submission follows on from our submission to the Budget Bill No 1.
Footnotes:
- Inspection générale des finances and Conseil général de l’environnement et du développement durable, Évaluation du dispositif d’aide fiscale à l’investissement locatif Pinel, November 2019: https://www.igf.finances.gouv.fr/files/live/sites/igf/files/contributed/Rapports%20de%20mission/2019/2019-M-036-05-Pinel_public-V2.pdf.
↩︎ - See most fully Cour des comptes, L’aide fiscale à l’investissement locatif Pinel, rapport d’évaluation de politique publique, September 2024, which also records the decision to let the scheme lapse on 31 December 2024: https://www.ccomptes.fr/fr/publications/laide-fiscale-linvestissement-locatif-pinel.
↩︎ - Income Tax Act 2007 (NZ), former subpart DH, in force from 1 October 2021, with the new-build exemption keyed to the date the code compliance certificate issued; repealed by the Taxation (Annual Rates for 2023–24, Multinational Tax, and Remedial Matters) Act 2024, with full deductibility of interest restored from 1 April 2025. ↩︎
- Stuart S Rosenthal, “Are Private Markets and Filtering a Viable Source of Low-Income Housing? Estimates from a ‘Repeat Income’ Model” (2014) 104(2) American Economic Review 687, estimating that rental housing filters down the income distribution at roughly 2.5 per cent a year in real terms: https://doi.org/10.1257/aer.104.2.687.
↩︎ - Assumptions used throughout: CPI 2.5 per cent per annum (indexation factor over eight years = 1.2184); marginal rate 47 per cent; incidental transaction costs ignored; holding periods treated as exact fractions of days held (precise day counts alter the figures immaterially); true 30 June 2027 values as stipulated. All figures rounded to the nearest $100.
↩︎ - HMRC, Capital Gains Manual, CG72602 — Assets held on 6 April 1965 disposed of with development value: https://www.gov.uk/hmrc-internal-manuals/capital-gains-manual/cg72602. The rule, originally in Finance Act 1965 (UK), sch 6, is now Taxation of Chargeable Gains Act 1992 (UK), sch 2, para 9(2).
↩︎
