Chapter 1 · Taxation and the lifetime earnings curve
The Peak
Earnings over a working life trace a curve: steep through your twenties and thirties, a peak around fifty, a slow descent to retirement. Australia's tax system takes its biggest bite exactly at the peak — which is exactly where the next hour of effort, the next promotion, the next business, gets decided.
The reason to care about marginal rates is that nobody decides whether to work at all. They decide whether to work more: take the overtime, chase the promotion, keep the business open another year, have the third kid and the bigger mortgage that comes with it. All of those decisions are priced at the margin. A worker at the peak of the curve, comparing another year of grind against time with their family, is doing arithmetic on the after-tax dollar whether they write it down or not. Price the marginal hour low enough and the answer is: don't bother. Multiply by ten million workers and thirty years. That is the quiet mechanism under a productivity chart that has gone flat.
A working life, before and after tax
Median full-time earnings by age (real), and what's left after income tax and the Medicare levy
Lifetime earnings (18–67)
–
gross, real dollars
Lifetime tax
–
income tax + Medicare
Kept of the next $1 at the peak
–
Peak-year average rate
–
of gross, that year
Assumptions & formula
data Earnings curve and tax schedule are sourced on the Assumptions page. The curve is pre-tax wages only — it excludes super, transfers, and family-payment tapers, all of which push effective marginal rates higher for parents in the middle of the curve, so this chart is conservative.
What you keep of the next dollar
Marginal retention by income — the price signal every extra hour of work actually faces
Steps are the 2025-26 brackets; the dip between ~$27k and ~$33k is the Medicare levy phase-in (10¢ per dollar until the full 2% catches up); the dashed line adds HELP repayments, which are a further 15–17¢ on every dollar over $67k until the debt clears. A graduate on $130,000 keeps 68¢ of their next dollar — 51¢ with a HELP debt, and 44¢ once they cross $135,000 with the debt still running.
Peak revenue is the wrong target
The Laffer curve is not controversial: at 0% taxation revenue is zero, at 100% nobody works and revenue is zero again, and somewhere in between sits a peak. The mistake, Casey argues, is treating that peak as the target. Growth peaks at a much lower tax share than revenue does, because by the time you have maximised this year's take, you have already consumed the economy's capacity to compound. A government that taxes to the revenue peak is optimising a single year and mortgaging every year after it. Run the horizon out a generation and the revenue-maximising rate slides left, towards the growth peak. Taxation past the point of peak growth is, definitionally, picking the pocket of your grandchildren.
The Laffer explorer
Casey's illustrative model: revenue now vs revenue compounded over a horizon you choose
Assumptions & formula
Casey asserts the functional forms and both peak locations — the Stagnation essay presents them as an illustrative model, not an estimate, and this page reproduces them faithfully rather than defending them. The measured anchor: data Australia's all-levels tax take is about 29–30% of GDP (ABS 5506). Casey's "roughly half" figure for the total burden counts a broader wedge than tax-to-GDP measures; the claims register scores that one.
The sector that stopped compounding
Here is the chart that should have opened the Productivity Summit. Since 2001 the market sector has lifted output per hour by about 40% — Casey's essay says 50%, the ABS series says 40; the shape is the argument. The non-market sector — public administration, health, education, the parts of the economy the taxpayer funds — is roughly where it sat in the mid-2000s, and health care specifically has gone backwards since 2010. Hold the level of service fixed and that gap is a bill: of the $839 billion Australian governments raised in 2024-25, Casey's arithmetic — run on the previous year's $800 billion — puts more than $250 billion, $10,000 per man, woman and child, down the hole dug by two decades of near-zero public-sector productivity growth. Rerun his arithmetic on the measured +40% gap and the bill is nearer $200 billion; the argument does not depend on the last fifty billion. A sector with no competition has no creative destruction, no price signals, and no mechanism to do more with less. It doesn't shrink when it underperforms. It bills.
Market vs non-market labour productivity
Output per hour worked, indexed to 2001 = 100
Source: ABS national accounts productivity estimates; see Assumptions for series details and the caveats on measuring non-market output (it is largely measured by inputs — which is itself the point).
Taxation hurts at least twice: once when it appropriates capital from the definitionally productive part of the economy, and again when it spends that capital employing scarce labour at productivity levels frozen in 2001. The essays' policy asks follow directly — fewer and simpler taxes, a public sector forced to find productivity the way everyone else does, and a tax share set for the decades, not the budget cycle. Chapter 3 runs the clock on what happens if we decline.
We could be the country where effort at the peak of the curve is worth keeping. The tax schedule is a policy choice, re-legislated whenever we like. Pick the horizon your grandchildren would pick.