I write the code behind the calculators on this site — the repayment engine, the tests, the pages. I went in thinking HECS arithmetic would be an afternoon's work: a threshold table and a multiplication. It wasn't, and most of what I had wrong turns out to be the same things the popular calculators have wrong. These are the five that changed the numbers the most.
Early on, my payoff projection had a bug I nearly shipped. I held the repayment thresholds indexed (they rise with average weekly earnings every 1 July) but left the default income growth at zero, thinking that was the conservative, assumption-free choice. Then I ran a perfectly ordinary case — $30,000 of debt on $80,000 — and the projection said the debt would never be repaid.
That output was arithmetically correct and completely misleading. If the thresholds rise 2.8% a year and your nominal income doesn't, you are taking a real pay cut every single year, and your repayment shrinks towards zero. "Income stays flat" isn't neutral — it is a strong, pessimistic forecast smuggled in as a default. Here is the same person under three assumptions:
| Assumption | Debt-free in | Total indexation paid |
|---|---|---|
| Income frozen at $80,000 forever | Not within 80 years | — |
| Income grows 2.8%/yr (same as thresholds) | 20 years | $10,386 |
| Income grows 3.5%/yr | 15 years | $8,099 |
$30,000 starting debt, $80,000 repayment income, 2026-27 rules, thresholds indexed at 2.8%/yr.
The gap between "never" and 15 years is not a detail — it is the whole answer. Any payoff tool that doesn't let you set income growth, or doesn't index the thresholds at all, is answering a different question than the one you asked. My own payoff grid states its assumptions in the first paragraph, because after that bug I don't trust any projection that hides them.
The single most common question about HECS — "why doesn't the amount coming out of my pay match any calculator?" — has a boring, structural answer: nothing that comes out of your pay is your repayment. Your employer withholds an estimate using the ATO's schedules; the actual compulsory repayment is worked out once a year from your annual repayment income, at assessment, and the two are reconciled then.
In 2025-26 that gap was systematic: the marginal repayment system started 1 July 2025 but the matching withholding tables only arrived on 24 September 2025, so anyone working through that window was over-withheld under the old, heavier schedules and gets the difference back at assessment. That refund is priced out here. Once I understood the withholding/assessment split, half the "calculator is wrong" complaints I'd read about other tools made sense — people were comparing a per-pay estimate to an annual formula and expecting them to agree mid-year.
This is the one I'd have got wrong on my own tax. Salary packaging reduces your taxable income, but HECS is charged on repayment income, which adds your reportable fringe benefits back — grossed up by 1.8868. Package $15,900 (a typical NFP cap) on a $90,000 salary and your taxable income falls, but your repayment income becomes $120,000: the packaged amount comes back as $30,000 of reportable fringe benefits.
Result: HECS of $7,571 instead of $3,071 — packaging added $4,500 a year to the HECS bill in that example. Packaging can still be worth it; the tax saving is real. But almost every packaging provider's marketing quotes the tax side and stays quiet about this side. The full table by salary and packaged amount is here.
Before the marginal system, HECS really did have cliffs — crossing a threshold by one dollar could add hundreds of dollars, because the old system charged a percentage of your entire income. That fear has outlived the system that caused it. I checked every dollar of income around the 2026-27 whole-of-income switchover: at $186,050 the repayment is $18,604.61, at $186,051 it is $18,605.10. The dreaded cliff costs 49 cents.
There's an even stranger detail at the other band boundary: because of how the 17% band's base amount is rounded, the repayment actually steps down by a few cents as you cross it. Nobody plans their salary around 18 cents, and now nobody should plan it around the "cliff" either. The full dollar-by-dollar check is here.
The reason old flat-rate calculators are still everywhere isn't laziness so much as silence: a HECS calculator that's wrong doesn't crash, it just returns a plausible wrong number, and no user can tell. The only defence I've found is mechanical: the engine behind this site runs a golden test suite — 32 assertions pinned to the ATO's published figures, sitting exactly on the band boundaries, the switchover point and both financial years — and the build fails if any number moves. When the ATO updates a threshold, the tests force me to look.
That's also why these guide pages are generated by the same engine rather than written with numbers typed in by hand: a table that's computed can't quietly disagree with the calculator next to it. If you want to check my working — or anyone else's — I've published the five test inputs I use, with the correct answers.
Every figure on this page is computed at build time by the same engine that powers the calculators on this site — nothing is typed in by hand, so the tables cannot drift away from the tools. The engine is checked against the ATO's published thresholds by a golden test suite that fails the build if any figure moves unexpectedly.