In short

For a CGT asset acquired before 21 September 1999, you can choose between the indexation method (frozen at the September 1999 quarter) and the 50% discount method, using whichever produces the lower taxable gain. Because indexation has been frozen for over 25 years, the discount now wins for almost every asset that has genuinely appreciated — indexation only wins for a flat or poorly performing asset.

For Australian retirees holding CGT assets acquired before 21 September 1999 — typically long-held investment properties, share parcels or business interests from the 1980s and 1990s — the tax law offers a choice of two ways to work out a capital gain on sale. The indexation method (Division 114 of the Income Tax Assessment Act 1997) uplifts the asset's cost base for inflation, but only up to the September 1999 quarter, after which indexation is frozen. The 50% CGT discount method (Division 115) instead halves the nominal gain for assets held more than 12 months. For an asset acquired before 21 September 1999, you may choose whichever produces the lower taxable gain. Most tax preparers default to the discount because it is simpler and usually wins — and, importantly, that default is now correct far more often than older guidance suggests, because indexation has been frozen for more than 25 years while asset values have kept rising. Still, for a genuinely low-growth asset the indexation method can win, so the comparison is worth running on each pre-1999 disposal.

The history explains the quirk. Before 21 September 1999, Australian CGT used indexation only: the cost base was lifted for inflation, and the residual real gain was taxed at marginal rates. The 1999 reform replaced indexation with the 50% discount for assets acquired from that date, and froze indexation at the September 1999 quarter for assets already held — no inflation uplift for any period after September 1999. That left pre-1999 assets with a choice that survives today, but with a steadily shrinking benefit: every year that passes adds nominal growth the indexation method cannot capture.

The indexation mechanics lift each cost base element by its own factor. The factor is the CPI for the September 1999 quarter — 68.7 — divided by the CPI for the quarter in which that cost was incurred, taken to three decimal places. Because the numerator is fixed at 68.7, the factors are modest and capped: an amount spent in the June 1991 quarter (CPI 59.0) carries a factor of 68.7 ÷ 59.0 = 1.164; an amount from the September 1985 quarter (CPI around 40.6) carries about 1.69, which is roughly the largest factor available; and an amount from the mid-1990s (CPI in the low-to-mid 60s) carries only about 1.05. The indexed cost base is subtracted from the sale proceeds, and the full indexed gain is taxable with no further discount. Note too that indexation can reduce a gain to nil but cannot create a loss — a capital loss is always worked out on the unindexed reduced cost base.

The 50% discount method is simpler: take the nominal gain (proceeds minus the unindexed cost base) and halve it, with the half taxable at marginal rates, provided the asset was held at least 12 months — a condition that is academic for a decades-old asset. The reason it usually wins in 2026 is arithmetic: halving a large nominal gain almost always beats lifting a relatively small cost base by a factor of at most about 1.7, especially when 25-plus years of post-1999 growth sits outside the indexation calculation. Indexation only wins where the asset's total nominal growth since acquisition has been small — that is, where the cost base, once indexed, approaches or exceeds the sale price. The further we get from 1999, the rarer that is.

A worked comparison makes the point. Take an investment property bought in the September 1990 quarter for $200,000 with $10,000 of buying costs, plus $50,000 of improvements in 1995, sold in 2026 for $1.5M. Under indexation, the $210,000 of 1990 costs lift by about 1.17 to roughly $246,000, and the $50,000 of 1995 improvements lift by about 1.05 to roughly $52,500, giving an indexed cost base near $298,000 and an indexed gain of about $1.2M, all taxable. Under the discount, the nominal cost base is $260,000, the nominal gain $1,240,000, and halving it gives $620,000 taxable. The discount wins comfortably — and by more than older "rules of thumb" assuming larger factors would suggest. Now contrast a genuinely flat asset: a property bought in the September 1985 quarter for $300,000 with $10,000 of costs, no improvements, sold in 2026 for just $500,000. Indexation lifts the $310,000 cost base by about 1.69 to roughly $524,000 — which exceeds the $500,000 proceeds, so the indexed gain is nil (and, since indexation can't create a loss, there is simply no taxable gain). Under the discount, the nominal gain of $190,000 halves to $95,000 taxable. Here indexation wins outright — but only because this asset barely grew in nominal terms over four decades, which is unusual.

The practical barrier is usually cost base reconstruction. Records from the 1980s and 1990s — purchase contracts, receipts for buying costs, improvement invoices — are often incomplete, and the indexation method needs a documented incurrence date for each element. Where records are thin, the ATO may accept reasonable estimates supported by contemporaneous evidence such as bank statements, but the effort can be significant. For retirees who expect to sell a pre-1999 asset, gathering that documentation early — while records still exist and memories are fresh — is far easier than reconstructing it at the point of sale. The timing of the disposal year itself also matters for the marginal rate the gain lands on.

What do worked planning examples show?

These two cases show how the indexation-versus-discount choice plays out in retirement disposals. Illustrative only — not personal advice — using FY25-26 rules.

Case 1 — Geoff, 68, retired, with an investment property bought in 1985 for $180,000, plus $80,000 of improvements added in 1995, now worth $1.1M. On these facts the comparison runs two cost base elements through indexation: the 1985 purchase lifts by about 1.69 to roughly $304,000, and the 1995 improvements lift by about 1.05 to roughly $84,000, for an indexed cost base near $388,000 and an indexed gain of about $712,000, all taxable. Under the 50% discount, the nominal cost base is $260,000, the nominal gain $840,000, and halving it gives $420,000 taxable. The discount wins clearly ($420,000 versus about $712,000), because the property's strong nominal growth swamps the capped 1999-frozen indexation. On these facts the rational step is to apply the 50% discount, while keeping the cost base records in case the ATO reviews the calculation.

Case 2 — Mary, 73, retired, holding a parcel of shares she bought in 1987 for $400,000 in a company that has languished, now worth only $470,000. This is the unusual case where indexation wins. The 1987 cost base lifts by roughly 1.5 to about $600,000, which exceeds the $470,000 sale price — so the indexed gain is nil, and there is no taxable gain (indexation having reduced it to zero rather than creating a loss). Under the 50% discount, the nominal gain of $70,000 would halve to $35,000 taxable. Indexation therefore eliminates the gain entirely here, purely because the holding barely rose in nominal terms over nearly 40 years. On these facts the rational step is to run the indexation calculation parcel by parcel — the laggards may favour indexation even though Mary's stronger-performing holdings will almost certainly favour the discount. The lesson is that the choice is made per asset (indeed per parcel), not once for the whole portfolio.

For retirees disposing of long-held assets, the indexation-versus-discount comparison is worth doing on every pre-1999 disposal — but with realistic expectations. In 2026, with indexation frozen at September 1999 and a quarter-century of growth sitting outside it, the 50% discount wins for almost all assets that have appreciated, often by a wide margin. Indexation now only wins for the genuine laggard — an asset whose nominal value has scarcely moved since acquisition. The advice work is to identify pre-1999 assets, reconstruct each cost base element with its incurrence date, apply the correct (and modest) CPI factors, compare both methods, take the lower-tax option, and document it. The default to the discount is usually right today — but the occasional flat or poorly-performing asset still rewards the check.

Sources

Key takeaways

  • For CGT assets acquired before 21 September 1999, you can choose whichever of the indexation method or the 50% discount method produces the lower taxable gain.
  • Indexation has been frozen at the September 1999 quarter since the 50% discount was introduced, so it can't capture any of the last 25-plus years of asset growth.
  • Indexation factors are modest and capped — even a cost incurred in September 1985 only gets a factor of about 1.69, the largest available.
  • Because of the freeze, the 50% discount now wins for almost every asset that has genuinely appreciated in value, and indexation typically only wins for an asset that has barely grown in nominal terms.
  • The method choice is made per asset (or even per parcel of shares), not once for a whole portfolio, so it's worth checking each disposal separately.

Frequently asked questions

Can I still use the indexation method for capital gains tax?

Yes, but only for CGT assets acquired before 21 September 1999, and only as an alternative to the 50% discount method — you choose whichever produces the lower taxable gain. Indexation is frozen at the September 1999 quarter, so it doesn't account for any growth in the asset's cost base since then.

Why does the 50% discount usually beat indexation for old assets now?

Because indexation has been frozen since 1999 while asset values have generally kept rising for the past 25-plus years, the indexed cost base only ever gets a modest uplift — at most around 1.69 times for the very earliest acquisitions. Halving a large nominal gain built up over decades of growth almost always beats that small, capped uplift.

When would the indexation method actually save more tax than the 50% discount?

When an asset has barely grown in nominal terms since it was bought — for example, a property or share parcel bought decades ago that's worth only modestly more today. In that case, indexing the cost base can bring it close to, or even above, the sale price, reducing the taxable gain to little or nothing, sometimes more than the 50% discount would achieve.

What records do I need to use the indexation method on an old asset?

You need the incurrence date and amount for each element of the cost base — the original purchase price, buying costs, and any capital improvements — since each is indexed separately using the CPI at the time it was incurred. Gathering this documentation (contracts, receipts, improvement invoices) well before you plan to sell is much easier than trying to reconstruct it at the point of sale.

A note on advice. This article is general information only and doesn't account for your personal circumstances. Everyone's situation is different — before acting, it's worth talking it through with a licensed adviser who knows your full picture.