← Back to “Against the CGT Discount”
Where every dollar of Australian CGT comes from
Master document. Single self-contained breakdown of Australia's ~$23B of CGT receipts in 2022-23 across 9 asset categories. Every number sourced, every method shown, every confidence rating spelled out. Updated May 2026.
Headline finding: About ~1% of every CGT dollar Australia raises comes from founders, early employees, and angel investors of fledgling Australian startups (companies with <100 employees, pre-Series B, <10 years old). The other 99% comes from property (~46%), listed public equities (~34%), mature private companies including Canva-tier unicorns (~5%), operating SMEs and family businesses (~4%), and a long tail of other assets. This finding is robust across multiple independent estimation methods — the range is 0.5%–1.2% in any plausible scenario.
The political debate around the 2026 CGT reform frames it as a threat to Australian innovation. But the dollar exposure of "true startup" gains is structurally tiny — partly because most startup exits happen at the scale-up tier or later, and partly because existing tax law (ESVCLP, ESIC) already exempts most genuine seed-stage innovation investment from CGT entirely. The reform's actual revenue comes from investors in mainstream asset classes — property, listed shares, family-held investments — held primarily by the top income decile.
What's in this document:
- The chart
- Summary table
- How trusts work in this analysis — and why "trust distributions" isn't a separate slice
- Methodology overview
- Per-slice breakdowns with multi-method triangulation
- 4.1 Investment property — 46%
- 4.2 Listed equities — 34%
- 4.3 Mature large private (Canva-tier + PE) — 5%
- 4.4 Operating SME / family company — 4%
- 4.5 Company operating-asset sales — 3%
- 4.6 Collectables & other — 3%
- 4.7 Crypto — 2%
- 4.8 Scale-up private (100-500 emp) — 2%
- 4.9 ★ True startup (<100 emp, <Series B) — 1%
- Honesty section — year choice, top-down anchoring, what could shift this
- Sources
1. The chart
Total CGT receipts ~$23B (2022-23 income year). Sorted by size. Trust-routed gains have been re-attributed to the underlying asset class (see Section 3) — so "Investment property" includes properties held in family discretionary trusts, "Listed equities" includes shares held via managed funds, etc.
2. Summary table
| Source of CGT | $M (2022-23) | % of total | Multi-method range | Confidence |
|---|
| Investment property (residential + commercial, all entities, includes trust-held) | ~$10,600 | ~46% | $8B – $14B | Medium-High |
| Listed equities (ASX + international, all entities, includes via managed funds) | ~$7,900 | ~34% | $6B – $10B | Medium-High |
| Mature large private (Canva-tier + PE-backed) | ~$1,150 | ~5% | $600M – $1.4B | Medium-High |
| Operating SME / family company shares | ~$900 | ~4% | $600M – $1.2B | Medium |
| Company operating-asset / business division sales | ~$800 | ~3% | $400M – $1.2B | Medium |
| Collectables & other (foreign assets, derivatives, etc.) | ~$670 | ~3% | $400M – $900M | Low |
| Scale-up private companies (100-500 employees) | ~$450 | ~2% | $250M – $650M | Low-Medium |
| Crypto assets | ~$400 | ~2% | $235M – $790M | Low-Medium |
| ★ True startup (<100 emp, <Series B): founder + early employee + angel gains | ~$200 | ~1% | $130M – $220M | Medium |
| Total | ~$23,000 | 100% | — | — |
3. How trusts work in this analysis
Why "Trust distributions" isn't a separate slice on the chart
Trusts in Australia are mostly flow-through entities for tax purposes. When a trust realises a capital gain, it almost always distributes that gain to a beneficiary (a person), who pays CGT on it at their marginal rate. Only a small portion of gains is retained at the trustee level.
The form vs. substance problem: When an individual receives a trust-distributed capital gain, they report it on the "Capital gain from a trust" line of their CGT schedule — not on the "Real estate," "Shares listed," or "Other shares" lines. So if I sliced the breakdown by tax-return line, I would have a sizable "Trust distributions" category that hides what's economically happening underneath.
This document attributes trust-distributed gains to the underlying asset instead. A family trust that holds an investment property and sells it appears in "Investment property" — not in a separate "Trust" slice. This is more economically honest and avoids the double-counting concern.
How the flow works
Trust owns investment property
↓
Trust sells property → realises $500K gain
↓
Trustee distributes $500K to family beneficiary
↓
Beneficiary reports $500K on tax return under
"Capital gain from a trust" CGT schedule line
↓
Beneficiary pays CGT at marginal rate (with 50% discount)
↓
In MY analysis: this $500K gain → "Investment property" slice
NOT a separate "Trust" slice
Estimated underlying composition of the ~$3.3B in trust-routed CGT
| Underlying asset | Share | $M | Reasoning |
|---|
| Investment property (family-held + A-REIT) | ~42% | ~$1,400 | Family discretionary trusts dominated by property; A-REIT distributions push this up |
| Listed Australian equities (direct + via funds) | ~29% | ~$960 | Family trusts often hold blue-chip portfolios; managed funds distribute equity gains |
| Private operating company / SME shares | ~15% | ~$500 | Family business interests typically held via discretionary trust for asset protection |
| International equities (via global funds) | ~9% | ~$300 | International fund distributions |
| Other (collectables, derivatives, etc.) | ~5% | ~$170 | Residual |
Why these shares are plausible
Per Treasury's discretionary trust fact sheet: "approximately 90 per cent of total private trust wealth is held by the wealthiest 10 per cent of households (those with net worth above around $2.3 million)." Per ACOSS via the Senate inquiry: "the highest 10 per cent of households ranked by wealth holding around 66 per cent of all investment property, shares, financial, and business assets." The composition above (heavy on property and family business + diversified portfolio of listed shares) matches what's known about Australian top-decile asset structures, where trusts predominate.
Double-counting check
The $23B total CGT is unchanged. We're redistributing the ~$3.3B that previously sat in a "trust distributions" line back to the underlying asset categories. This makes "Investment property" larger (from ~$9.2B to ~$10.6B) and "Listed equities" larger (from ~$6.6B to ~$7.9B), while eliminating the separate "Trust" slice. Each dollar is counted exactly once.
Verdict on this attribution: medium confidence. The 42/29/15/9/5 split is reasoned from Treasury's wealth-distribution data, ACOSS analysis, and ABS household composition, but is not directly published. If the split were materially different (e.g., 30% property / 40% equities), Property would be ~44% of CGT (instead of ~46%) and Listed equities ~36% (instead of ~34%) — the headline narrative is robust to reasonable variation.
4. Methodology overview
Step 1 — Top-down: total CGT and entity split
Total Australian CGT receipts for 2022-23 ≈ $23 billion, per Treasury reporting and ATO Taxation Statistics 2022-23. Capital gains aren't taxed as a separate tax in Australia — they're a component of income tax. "CGT receipts" is calculated via the ATO's estimated tax on net capital gains method: net tax ÷ taxable income × net capital gain, applied per entity.
The ATO's published Chart 16 splits this across entities:
- Individuals (including via trust distributions): ~$18-19B (~80-83%)
- Super funds: ~$2-3B (~10-13%)
- Companies: ~$1-2B (~5-8%) — taxed at corporate rate; no discount
- Trusts (retained, not distributed): ~$0.5B (~2%)
Step 2 — Within each entity, apportion to underlying asset classes
The ATO publishes CGT schedule line items (Real estate, Shares listed, Other shares, Units, Collectables, etc.) only at aggregate level — the underlying dollar split isn't separately published. This step uses triangulation: ABS household wealth composition, Treasury TEIS distributional analysis, Senate inquiry findings, ATO rental property statistics, AIC/Cut Through Venture industry data, and international comparison (US IRS).
Step 3 — Re-attribute trust-routed gains to underlying assets
See Section 3 above. The ~$3.3B that flows through the "Capital gain from a trust" tax-return line is redistributed to the underlying asset categories it economically represents.
Step 4 — Triangulate each slice from independent angles
For each slice I built 2-4 independent estimates using different data sources and reported whether the methods converge. Per-slice breakdowns below show each estimate explicitly with the math walked through.
5. Per-slice breakdowns
5.1 Investment property — ~$10,600M (~46% of CGT)
Includes residential and commercial investment property held directly by individuals, through trusts, by super funds, and by companies. Owner-occupied housing is excluded (main residence exemption — see honesty section).
Estimate A — Top-down with trust re-attribution
Direct individual property CGT:
45% × $19B = $8.55B
Plus trust-routed property CGT (42% of $3.3B trust slice):
$1.4B
Plus super-held commercial property:
$0.4B
Plus company-held property:
$0.5B (corporate property gains taxed at company level)
Plus trustee-retained property CGT:
$0.1B
Total: ~$10.95B → ~$10.6B (rounded for shared margin of error)
Estimate B — Transaction volume × investor share × gain × effective rate
Australian property transactions in 2022 =
$674.5B across 730,000 settlements. 2023 was down 9.1%. Average 2022-23 fiscal year ≈ $640B in transactions. Investor share of new lending: 33.5% in 2022. Investor-side activity (buy + sell) ≈ $214B. Assuming 8-10 year holding period and ~6% p.a. appreciation, gain fraction at sale ≈ 40-50% of sale value.
$214B × 0.45 = $96B gross gains → × 50% discount → $48B taxable → × ~28% weighted marginal rate →
~$13B CGT
Estimate C — From Treasury TEIS CGT discount cost × property share
CGT discount for individuals/trusts in 2022-23 cost
$23.46B revenue forgone (Treasury TEIS 2025-26). Senate inquiry: housing is "substantial share." Empirical and Treasury framing suggests ~40-50% of discount benefit attaches to property. Property's discount cost ≈ $9-12B. Since the discount halves the tax, actual receipts ≈ similar order =
$9-12B.
Verdict: strong convergence — three methods give range $8-14B; headline $10.6B sits in the middle. Trust the rank and the rough magnitude; precise dollars could vary ±20%.
5.2 Listed equities (ASX + international) — ~$7,900M (~34% of CGT)
Includes shares held directly by individuals, through trusts and managed funds, by super funds, and by companies. Includes both Australian-listed (ASX) and international shares held by Australian residents.
Estimate A — Top-down with trust re-attribution
Individual direct listed Australian shares:
18% × $19B = $3.42B
Individual direct international equities:
6% × $19B = $1.14B
Plus trust-routed listed Aus equities (29% of $3.3B × ~70% Aus share = ~20%):
$0.66B
Plus trust-routed international equities (9% of $3.3B):
$0.30B
Plus super fund equity gains (~80% of $2.5B super CGT):
$2.0B
Plus company-held listed share investments:
$0.2B
Plus trust-retained listed CGT:
$0.2B
Total: ~$7.92B
Sources: derived from ATO entity totals + APRA super fund asset composition + trust attribution.
Estimate B — Bottom-up from holdings × turnover
ABS Sep 2024: households hold
$1.46T directly in listed equities; super funds hold ~$4T (~50% in equities ≈ $2T). Annual turnover (sales) typically 15-25% of holdings = $260-490B sold per year. Average gain fraction ~30-40%. Gross gains ~$80-200B. Super CGT taxed at 10% effective; individual direct portfolio gains heavily skewed to top decile. Recovered CGT ≈
$6-10B.
Estimate C — International benchmark (US IRS scaled)
US IRS SOI: corporate stock gains =
18.5% of all US individual capital gains. Adjusting for Australian differences (lower direct household equity ownership but more accumulated via super), expected Australian listed-shares share of individual CGT after trust re-attribution:
~25-35%. My estimate (24% direct + 6% trust-routed = 30%) lands in the middle.
Verdict: strong convergence — three methods give $6-10B; headline $7.9B at the upper-middle. One of the most trustworthy numbers in the analysis.
5.3 Mature large private (Canva-tier + PE-backed) — ~$1,150M (~5% of CGT)
Includes founders, executives, and ESS holders of large unlisted Australian companies (Canva, Airwallex, Employment Hero etc.) plus PE-backed buyouts of mid/large private operating businesses. Defined as 500+ employees OR $1B+ valuation OR established 10+ years.
Estimate A — Bottom-up from PE buyout volume
2022 Australian PE buyout deal value:
A$57-66B. 35% × $57B ≈ $20B in private-private deals (excluding take-privates of listed). Individual seller share ~40-60% ≈ $10B accruing to individuals. Gain fraction ~70% ≈ $7B gross. Apportioned across 2022-23 fiscal year (straddling 2022 boom + 2023 decline): ~$4-5B gross. ~60% Tier 3: ~$2.4-3B gross Tier 3. After 50% discount × 47% top marginal =
$560-700M. Plus unicorn secondaries ~$300-400M.
Subtotal: ~$850M-$1.1B. Plus trust-routed mature-private holdings (~$150M).
Total: ~$1.0-1.25B
Estimate B — Named-deal sum
2022-23 transactions identifiable in trade press:
- Canva: Blackbird sold $150M at $39B valuation Aug 2023. Internal employee tenders FY22-23 (smaller). Canva-related gross gains 2022-23: $200-400M.
- Airwallex: Series E extension Oct 2022 at $5.6B. Modest secondary: $50-150M.
- Other unicorn activity (Culture Amp, Linktree, Employment Hero): $50-150M.
- Large PE buyouts of established mid/large private operating companies (~$20-25B of the $57B 2022 PE buyout value): individual seller gross gains: $600-900M.
Named transactions:
$900M-$1.6B gross → ~20% effective rate =
$180-320M CGT from named deals alone. Real number higher; trade press misses transactions.
Estimate C — AFR Rich List wealth turnover
AFR Rich List 2024: top 200 =
$625B combined wealth; tech sector $108B. Annual turnover ~3% of wealth ≈ $19B in realised transactions. Gain fraction ~70% (low cost basis). ~30-40% in Tier 3-qualifying companies: ~$4-5B gross gains. After CGT × 47% top marginal:
$600M-$1.2B from Rich-List-tier alone.
Verdict: strong convergence — three methods give $600M-$1.4B; headline $1.15B well-anchored.
Important note for the blog: Canva's biggest secondaries were 2024-25 (the $1B tender at $26B and the $42B oversubscribed tender), not 2022-23. In our base year Canva was in a down round. So Tier 3 in 2022-23 is driven more by PE-backed mid/large private buyouts than tech unicorn secondaries. If you ran the analysis on 2024-25 data this slice would be substantially larger — the "Canva matters way more than scrappy startups" message would be even stronger.
5.4 Operating SME / family company shares — ~$900M (~4% of CGT)
Individuals selling shares in non-VC-backed private operating companies — family businesses, professional services partnerships, manufacturing SMEs, trade businesses. Many use small business CGT concessions (preserved under the 2026 reform).
Estimate A — Via small business CGT concession claim volumes
Treasury TEIS 2025-26 reports small business CGT concession revenue forgone:
- E32 — Small business 50% reduction: $1,180M
- C12 — 15-year exemption: $1,190M
- C7 — Retirement exemption: $830M
- Total: ~$3.2B revenue forgone
Implies underlying gross gains of ~$10-15B where concessions apply. Of those, ~30-40% are share-based (rest is goodwill, real estate, plant). After discounts and marginal rates:
~$200-300M from concession-claiming SME share sales. Plus non-concession SME sales: ~$300-400M.
Subtotal:
~$500-700M. Plus trust-routed SME shares (much of family business held via discretionary trust):
~$300M.
Total: ~$800M-$1.0B
Verdict: medium confidence — share/goodwill split is fuzzy; trust attribution adds material to this slice. Range $600M-$1.2B. Trust the order of magnitude; precise number depends on assumed share/goodwill split.
5.5 Company operating-asset / business division sales — ~$800M (~3% of CGT)
CGT paid by companies (the legal entity itself) when the company sells business assets — divisions, plant, intellectual property, goodwill on partial sale. Companies pay at the corporate rate (25%/30%) with no discount. NOT the same as individuals selling shares in a company.
Estimate A — Top-down within company CGT
Companies pay ~$1.6B in CGT receipts in 2022-23 (corporate rate; no discount). Within this:
- Operating asset / business division sales (plant, equipment, IP, goodwill): ~50% → $800M
- Property holdings: ~30% → $500M (already in property slice)
- Investments in other entities: ~15% → $200M
- Other: ~5% → $100M
Real-world examples: a conglomerate divesting a non-core business; a manufacturer selling its factory equipment and IP; a private parent company selling a subsidiary.
Source: ATO company tax statistics; top-down estimate (companies' CGT isn't separately published).
Verdict: medium confidence — companies' CGT isn't separately published from ordinary company tax. The $800M could be $400M-$1.2B. Trust the direction (companies are a small CGT contributor); precise number is fuzzy.
5.6 Collectables & other — ~$670M (~3% of CGT)
Catch-all for assets not elsewhere captured: art, jewellery, antiques, classic cars; foreign-sourced gains not in international equities; derivatives, options; trust-retained CGT not attributable to a specific asset category.
Estimate A — Residual + small components
Sub-components (rough):
- Collectables proper (art, jewellery, antiques): ~$50-100M
- Foreign-sourced gains not in international equities (rental property overseas, foreign-incorporated entities): ~$150-250M
- Trust-retained CGT not in other categories: ~$200-300M
- Other CGT events (options, derivatives, business rights): ~$50-100M
- Trust-routed "other" (5% of $3.3B): ~$170M
Total: ~$600-800M
Source: no good independent source — constructed as residual + sub-components.
Verdict: low confidence — this slice is the catch-all. Range $400M-$900M. Don't quote precise numbers from this slice in the blog; just note it's small.
5.7 Scale-up private companies (100-500 employees) — ~$450M (~2% of CGT)
Founders, executives, and employees of growth-stage Australian companies past Tier 1 (true startup) but pre-unicorn. Examples: SafetyCulture around Series C, Linktree before $1B+ valuation, Octopus Deploy.
Estimate A — Census × event frequency + trust attribution
Airtree census: ~100-150 active Australian scale-ups in this tier. ~15% have meaningful liquidity event in any year = ~18 events × ~$30M individual gains = $540M gross × 50% discount × 40% marginal =
$108M.
Plus M&A acquisitions (~$70M CGT), ESS exercises (~$50M), founder secondaries at non-exiting scale-ups (~$60M), late-stage tender offers (~$40M), minority PE recaps (~$70M). Component sum:
~$400M.
Plus trust-routed scale-up holdings: ~$50M.
Total: ~$450M
Verdict: low-medium confidence — last 3 components don't have hard transaction counts. Range $250M-$650M. The rank (scale-up > true startup, scale-up < mature private) is robust; precise number isn't.
5.8 Crypto assets — ~$400M (~2% of CGT)
Australian-resident individuals realising gains on crypto holdings. Companies and super funds hold trivial amounts; SMSF crypto holdings are ~$1.7B with minimal annual gains.
Estimate A — Top-down from asset pool
Australian crypto holdings ≈
$22B (Independent Reserve / Roy Morgan surveys, 2024). 25% annual turnover × 30-40% gain fraction = $1.6-2.2B gross gains. After 50% discount × 35% weighted marginal =
$280-385M CGT.
Estimate B — Bottom-up from declared holders
~600,000 individuals have declared crypto on tax returns (ATO via data-matching). If avg declared gain per declarer ≈ $5-10K/year, only ~30% report positive net gain in a year: gross gains ~$1.0-1.8B. After discount and marginal:
~$200-300M CGT.
Verdict: low-medium confidence — range $235M-$790M. ATO doesn't publish crypto CGT separately; compliance is patchy. Headline $400M is the midpoint of a fuzzy range; ±50%.
5.9 ★ True startup (<100 employees, <Series B) — ~$200M (~1% of CGT)
This is the headline finding. Founders, early employees, and angel investors of fledgling Australian startups — the cohort the political "protect Aussie innovation" debate is supposedly about. Four independent methods converge on the same answer: ~1% of total CGT.
Estimate A — Bottom-up from Australian startup exits
~40-80 successful Tier 1 exits per year. Average exit value ~$20M. Of consideration, ~60% to founders + early employees + angels (rest to VC funds and institutional).
60 events × $20M × 60% = $720M individual gross → 80% is "gain" = $576M → 50% discount → $288M taxable → 37% weighted marginal =
~$107M.
Plus secondaries and ESS at Tier 1:
+$50-100M.
Total: ~$150-200M. (Trust attribution ~$0M — seed-stage rarely uses family trust structure.)
Estimate B — Structural: ESVCLP and ESIC already exempt most of this tier
The most important argument. Existing tax law provides full CGT exemption for the politically-salient innovation investor cohort:
- ESVCLP (Early Stage Venture Capital Limited Partnerships): per ATO, "Australian resident and foreign resident limited partners are exempt from income tax on revenue and capital gains from disposals of eligible venture capital investments made by the ESVCLP." Any gain through an ESVCLP-eligible vehicle is fully CGT-exempt.
- ESIC (Early Stage Innovation Company): full CGT exemption for qualifying investments held 1-10 years, plus 20% non-refundable tax offset on the investment itself.
Implication: the CGT we currently collect from "true startup" gains is by definition the
residual — gains that
didn't qualify for ESVCLP or ESIC. That residual is small
by construction.
The carve-out for genuine innovation investment already exists in law.
Estimate C — Cross-check via Australian VC capital flow
Australian VC deployed $4.0B across 414 deals in 2024 (up 11% from 2023). Most goes to companies of Tier 1-2 scale. With 3x avg return on successful investments and 20-30% success rate, annual VC-track "exit" returns ~$0.5-1B/year for the industry. Bulk flows to LPs (super funds, ESVCLP-eligible, institutional). Flow to individual founders + early employees + non-fund angels: ~$100-300M gross → after CGT:
~$30-80M from "pure" Tier 1 founder/employee gains. Plus boundary cases with Tier 2: ~$100M.
Total: ~$130-180M.
Estimate D — Comparison to other small slices
Sanity check: "true startup" should be similar in scale to other narrow asset classes. Crypto ~$400M (~600K declared users), collectables/other ~$670M. True startup at ~$200M means it's smaller than crypto and smaller than collectables/other. Given ~2,000-3,000 active Tier 1 Australian startups and ~50-100 successful exits/year, with most successful Tier 1 founders using ESIC-qualifying structures, the small absolute number is plausible.
Cross-comparison.
Verdict: strong convergence across all four methods — range $130M-$220M. The "true startup is tiny" finding is among the most robust in the entire analysis.
Across all plausible parameter values, true startup CGT lies between 0.5% and 1.2% of total CGT. This is one of the most trustworthy numbers in the analysis — three methods are simple bottom-up calculations, and the structural argument (ESVCLP/ESIC carving out the politically-salient cases) is unambiguous law.
6. Honesty section
6.1 Which year, and why 2022-23
2022-23 is the most recent year for which ATO has published comprehensive Taxation Statistics (released 27 June 2025). The 2023-24 stats are due around June 2026 — very soon, but not yet available. This is the genuine data frontier, not a comfortable old year.
However, 2022-23 has two structural quirks worth flagging:
- Canva was in a down round. Valuation cut from $40B to $26B in 2022. The big secondaries (the $1B tender at $26B in 2024; the $1B oversubscribed at $42B in 2025) hadn't happened yet. So the "Mature Private" tier is depressed vs. what 2024-25 will look like.
- PE buyout activity was at a transition point. 2022 was a record year (~A$66B); 2023 dropped to ~A$13B. Our fiscal year straddles both.
Implication for the blog: the "true startup is tiny" finding is robust to year choice. If anything, it gets stronger in more recent years because Mature Private grows (Canva-driven) while True Startup stays roughly constant.
6.2 Top-down vs bottom-up: where the analysis is anchored
This is a top-down constrained analysis. The total CGT base (~$23B) is the most reliable single fact; my job was to apportion it across 9 asset categories. Where bottom-up component estimates didn't sum to the top-down total, I revised the components. This means:
- The shares (percentages) are more trustworthy than absolute dollars on individual slices.
- The rank ordering is very robust.
- The tightness of granular sub-slice estimates (scale-up, crypto) is moderate — they're partially fit to the residual.
6.3 What could shift this picture meaningfully
- ATO sample file or detailed CGT schedule asset-class breakdown. Would let us replace the asset-class apportionment within individual CGT with real numbers. Biggest possible improvement.
- 2023-24 or 2024-25 ATO stats. Would let us see the Canva mega-secondary period directly. Likely strengthens "Tier 3 mature private dominates True Startup."
- PBO costing internal asset-class breakdown. Almost certainly has a sharper version of this; not publicly released but potentially FOI-able.
- APRA super fund asset-class CGT data. Would refine the super fund split (currently estimated 80% equities / 15% property / 5% other).
- Better trust composition data. The 42/29/15/9/5 split for trust-held assets is reasoned but not directly published. ATO trust statistics + ABS Household Wealth and Income Survey microdata could tighten this.
6.4 What the analysis can and can't support
Strong claims:
- Property + listed equities account for ~80% of Australian CGT receipts (was 69% before trust re-attribution, now 80% with it). Robust across all methods.
- True startup founder/employee gains account for ~1% of CGT receipts (range 0.5-1.2%). Robust across four independent methods including the structural ESVCLP/ESIC argument.
- The carve-outs to protect Australian innovation already exist in current law (ESVCLP, ESIC). The 2026 reform's "consultation on early-stage investment" is fine-tuning of an existing framework, not creating a new exemption.
- The mature large private slice is several times larger than the true startup slice — and the gap widens in 2024-25 and 2025-26 data.
Weaker claims (data only roughly supports):
- Precise dollar amounts of crypto or collectables CGT — these slices are genuinely fuzzy.
- Exact split between scale-up (Tier 2) and operating SME (Tier 4). Rank robust; precise numbers aren't.
- Forward projection to 2024-25 or 2025-26 without explicit modelling of Canva activity.
6.5 The owner-occupied housing slice that's intentionally absent
The main residence exemption means owner-occupiers pay zero CGT despite holding ~$8.5T in housing equity. Treasury TEIS estimates the revenue forgone from this exemption at ~$60B/year — nearly 3× the entire CGT base. If you included "revenue forgone via main residence exemption" the entire chart would be dominated by that one line. This document is "what we do collect," not "what we choose not to collect." But it's worth knowing the scale of the carve-out.
7. Sources
Primary government sources
Industry data
Specific transactions and policy
International comparison