Every one of these was found by checking the model against a source document, an enterprise agreement, or a serving member. They are listed because a tool you cannot audit is a tool you should not trust.
Note which way they run. Five of these corrections reduced the defined benefit or raised the accumulation figure; two raised the defined benefit by removing money the accumulation side was never entitled to. The corrections cut both ways, which is the only evidence worth offering that the model was not steered toward a conclusion. Anyone reviewing this work should check that claim rather than accept it.
The default investment return was below every period in the historical record. The model used the scheme actuary’s 8% gross assumption, which after the 15% earnings tax and investment fees came to 6.27% net — lower than the worst 25-year stretch balanced options have ever delivered, and about 1.3 points below their long-run average. Because a defined benefit does not move with markets, the return assumption only ever changes the accumulation side, so an over-conservative figure is not neutral: it silently penalises the option it is applied to. The default is now the median return balanced options actually returned over a career of the length being modelled, net of fees and tax, with the full historical range marked on the slider. This correction favours the accumulation side, and materially: on the standard comparison it widened the accumulation lead from about $94,000 to about $265,000. It is listed here in that plain form because a change of that size, running in the direction of the model’s own argument, is exactly the kind a reader should scrutinise hardest.
The Victoria Police rank bands were mislabelled. Leading Senior Constable was shown as beginning at increment 10; it begins at 13. Senior Constable runs 5–12, and increment 3 is First Constable, not Constable. The pay rates and the salary path were correct throughout — only the labels were wrong — so no figure changed. Corrected against a serving member.
Post-cap contributions were overstated. The model was redirecting employer and member contributions into the post-cap account. ESSSuper’s FS034 says members are required to stop contributing at the maximum multiple — only the employer keeps paying. The tool was crediting the defined benefit with money members never actually pay. Fixed: employer only.
AV’s 2016–2019 pay was 20.7% too low. Those years were interpolated smoothly between the 2015 and 2020 agreements — straight through the 2015 EA Work Value Case, whose third increase landed 1 December 2016 and lifted rates about 20%. Fixed: re-anchored to the published rates, all 15 classifications now match to the dollar.
The scheme's death and disability benefits were being modelled as insurance. The accumulation side was charged 0.6% of salary in premiums to “match” cover the scheme supposedly gave away free. It gives away no such thing. These benefits are projected to age 55 — worth a great deal at 25, very little at 54, and nothing at all from 55. That is the opposite shape to a policy that pays a fixed sum insured on top of a balance at any age, and the two cannot be priced against each other. Fixed: removed from both sides. This correction runs against the defined benefit — it returned about $40,000 to the accumulation side.
The accumulation side was credited with employer money members never receive. The model used each employer’s scheme funding rate on the accumulation side. But what an employer pays to keep the fund solvent (currently 16% for Victoria Police, 14.5% for Ambulance Victoria, 12% for Fire Rescue Victoria) is not a member entitlement — the benefit formula is multiple × final average salary and does not contain the employer rate at all. Crediting an accumulation fund with 14.5% handed members money they would never have got, most of it in the early years that compound longest. Fixed: the accumulation side now receives the Superannuation Guarantee, the same for every service, because that is what an employer would actually have to pay into an accumulation fund. This correction favours the defined benefit — it cut the accumulation side by roughly $50,000–$60,000 and reversed the result for Ambulance Victoria.
The redirected account was missing its contributions tax. When the post-cap fix above was made, the account stopped applying the 15% contributions tax that every other concessional dollar in this model pays. It was overstating the redirected account by 15%. Fixed.
The money you stop contributing was unaccounted for. After the correction above, the member’s own contribution simply vanished from the model — it was neither invested nor acknowledged, even though on a long career it is well over $100,000 of the member’s own money. It is now an explicit choice, because it is an assumption and not a fact.
Fees defaulted to zero. No fund is free, and zero flattered the accumulation side. Fixed: the default is now AustralianSuper’s MySuper Balanced published rates, which reproduce their own quoted $367 a year on a $50,000 balance.