Super calculator

Each fund offers several investment options, each with its own fee. Compared here is each fund's option closest to 100% shares. Every fund lists three kinds of cost in its Product Disclosure Statement (PDS):

• Administration fees and costs. A flat fee (Hostplus: $78 a year), a percentage of your balance (Cbus: 0.19%, up to a dollar cap in some funds), and costs paid from the fund's administration reserve (Hostplus: $51.28 per member, Cbus: 0.04%). Reserve costs don't leave your account but are paid from members' money, so the disclosure rules (ASIC RG 97) count them, and so does this page.
• Investment fees and costs. About 0.02–0.05% for an index fund, 0.3–0.6% for an actively managed one.
• Transaction costs. Near zero for index options.

At $63,000, 50% Australian:

• Hostplus: $\underbrace{\$78}_{\text{flat fee}} + \underbrace{\$51.28}_{\text{from reserve}} + \underbrace{0.025\% \times \$63{,}000}_{\text{investment}} = \$145$
• Cbus: $\underbrace{\$52}_{\text{flat fee}} + \underbrace{0.190\% \times \$63{,}000}_{\text{admin \%}} + \underbrace{0.040\% \times \$63{,}000}_{\text{from reserve}} + \underbrace{0.415\% \times \$63{,}000}_{\text{investment}} = \$458$

In today's dollars, at a 6% real return (after inflation) in every fund, so only fees differ. Each year:

Hostplus, year one: $\$63{,}000 \times 1.06 + \$7{,}548 - \$145 = \$74{,}183$. Fees compound too: a dollar paid at 44 would have been $1.06^{20} \approx \$3.21$ by 64.

Your employer pays the Superannuation Guarantee, 12% of salary: $\$8{,}880$. After 15% contributions tax, $\$7{,}548$ is invested.

No co-contribution above $64,293.

Each fund's default option (where new members go) over 10 years, from APRA's yearly performance test. The test compares a fund's return with an index version of the same mix: the same split across shares, bonds and property, each earning an index return. Above zero, the fund's choices beat their extra cost. Too far below, it fails.

Default options hold different amounts of growth assets, which explains much of the gap in raw returns. Unlisted assets are revalued only every few months, so they look smoother than they are.

Every diversified option APRA tests for these funds. Right is more growth assets, up is a higher 10-year return. Returns are after investment fees and tax, before administration fees; cost includes both at $50,000. Options under 10 years old are in the table only.

Call the Australian share of your money $w$. Australia is about 2% of world shares, so even 50% leans home. Three things set the best $w$.

1. Average return

Over the 30 years to 2026, Australian shares compounded 9.0% a year and international 8.6% (Vanguard), so treat them as equal. The model uses the average yearly return $\mu$, which is higher than the compound rate (see below): $\mu = 10.1\%$ for both. A mix's average:

$A$ is Australian, $I$ international, $p$ your portfolio. With equal averages, only franking and risk separate them.

2. Franking credits

Australian companies pay 30% company tax and pass shareholders a franking credit for it. A company earning $100 pays $30 tax and you $70 plus a $30 credit. Super owes 15% on the $100, so the ATO refunds the other $15. Overseas tax earns no credit.

Credits on the All Ordinaries are worth 1.04% a year (ATO, September 2026). After super's 15% tax that's $1.04\% \times 0.85 = 0.88\%$ extra a year on Australian shares: the franking benefit.

3. Risk

Volatility $\sigma$ is the standard deviation of yearly returns. International returns include currency risk, which often helps: the Australian dollar tends to fall in crashes. Over 2000–2025, $\sigma_A = 15.4\%$, $\sigma_I = 17.0\%$ and their correlation $\rho = 0.58$ (−1 opposite, 1 together). A mix's volatility:

At $w = 0.50$ that's 14.4%, against 15.4% for Australia alone and 17.0% for international alone. Mixing helps because $\rho < 1$.

Why risk costs you

Up 50% then down 50% turns $100 into $75, despite a 0% average. What you end up with grows at the compound rate $g$, which volatility drags below the average:

At your split with no franking benefit: $10.1\% - \tfrac{(14.4\%)^2}{2} = 9.08\%$. The least volatile mix grows most: 61% Australian with no franking benefit, 100% with the full benefit.

Sharpe ratio

The Sharpe ratio is return above cash per unit of volatility:

with $r_f$ the cash rate, 4.85% (RBA, 4 November 2026). It picks the best mix assuming you can borrow to add risk. Super can't borrow, and with 25 years to wait you want all shares anyway, so the target is the highest $g$. With no franking benefit both pick 61%; with the full benefit Sharpe picks 76% and $g$ picks 100%.

The unknown

If franking is already in Australian share prices, the benefit is smaller. It's somewhere between 0 and 0.88%. The regret of a mix is the growth it gives up against the best mix for that benefit:

Across 11 benefit values, the mix with the smallest worst-case regret (minimax regret) is 81% Australian, giving up at most 0.05% a year.

Why 50%

The model sees 26 years and assumes a bell curve, so it can't see a long national slump: Japan's 1989 peak wasn't passed until 2024. The ASX is also about half banks and miners. So 50%, a judgement call, which costs at most 0.29% a year in the model.

$10,000 from 2000, rebalanced to the split each year, before fees and franking. Per year is $\left(\tfrac{\text{end value}}{\text{start value}}\right)^{1/\text{years}} - 1$; max drawdown is the largest fall from a previous high, at year-ends. International returns are in Australian dollars:

The dollar's fall from about US$1.05 to US$0.65 after 2012 is part of why international did so well.