Why raising tax rates doesn't raise revenue in proportion
Taxpayers respond to higher rates by changing behaviour, working less at the margin, restructuring how income is received, deferring or shifting assets, sometimes relocating, and that behavioural response means the revenue gain from a higher rate is consistently smaller than static arithmetic predic
Raising a tax rate and raising tax revenue sound like the same action described two ways. The evidence says they are not, and the gap between them is one of the more consistently underestimated dynamics in fiscal policy debate.
The shape of the relationship
At a 0 percent tax rate, government collects no revenue, for the obvious reason that there is no tax to collect. At a 100 percent rate, government also collects no revenue, because nobody works or invests when the entire return is confiscated. Somewhere between those two points sits the rate that maximises revenue, and the genuinely interesting empirical question, one this piece doesn't claim to answer for every bracket and every country, is where that point sits and how revenue responds as rates move around it. The concept itself is not controversial. The political argument is entirely about where current rates sit relative to that point, and what taxpayers actually do in response when rates change.
What the data shows when it's actually measured
A detailed study of the Spanish income tax system found that behavioural responses to taxation cause a loss of 53.77 percent of the revenue a rate increase would otherwise be projected to raise, meaning roughly half of any projected gain gets absorbed by changed taxpayer behaviour before it reaches the treasury. The same study found around 44.72 percent of Spanish taxpayers sit on the "normal" side of the curve, where higher rates would still raise more revenue from them, with the remainder at or past the revenue-maximising point for their bracket. A major American study covering six decades of tax reform from the 1920s onward, led by Austan Goolsbee, found no historical evidence the US has been positioned to the right of the revenue-maximising rate, suggesting room existed to raise top rates. But the same study found the Laffer curve is considerably flatter than the popular graphic suggests, because revenue changes from raising or cutting top rates are muted by behavioural change. The critical finding was that even when labour supply doesn't shift at all, income shifting, moving income out of taxable salary into deferred compensation, capital gains, corporate structures or in-kind benefits, substantially erodes the revenue a rate increase was projected to raise.
The response that matters most isn't working less
The most underappreciated behavioural response isn't "high earners work fewer hours." It's that high earners restructure how they receive income in the first place. When marginal rates rise, salary gets reclassified as capital gains, taxed at a lower rate. Income gets deferred into superannuation or corporate structures. Trusts and family companies distribute income to lower-bracket family members. The timing of asset sales and income recognition shifts across tax years to land in a more favourable period. High earners with genuine international mobility relocate entirely. None of these responses require reducing actual economic activity, they simply move income into lower-taxed or deferred forms, leaving government with less revenue than the static projection promised, and an economy that has spent real resources on tax minimisation rather than anything productive.
Not all income responds the same way
The size of the behavioural response depends heavily on the type of income involved. Wage income is relatively inelastic, workers don't dramatically cut their hours in response to a tax rise, though high earners retain more flexibility than low earners to adjust in other ways. Capital income is highly elastic by comparison, investors can time when they realise gains, hold assets longer, shift jurisdictions or restructure to defer tax entirely. High-income earners sit at the intersection of both effects, facing the highest marginal rates while having the most resources and options to respond to them, which is exactly why revenue gains from raising the top bracket tend to disappoint static projections more than gains from broader-based tax changes.
Australia's version of the debate
Australia's top marginal rate is 45 percent, 47 percent including the Medicare levy, applying to income above $180,000, alongside a 30 percent company tax rate, 25 percent for small business, and capital gains taxed at marginal rates with a 50 percent discount for assets held over twelve months. Treasury's own revenue costings already build in behavioural adjustments when modelling proposed changes, an implicit acknowledgment of the principle even where the precise magnitude remains debated. Australia's imputation system for franking credits adds a further layer of complexity to any simple rate comparison, since it changes the effective tax rate investors actually face on dividend income. Recurring debates over the top marginal rate, superannuation tax concessions, the capital gains discount and negative gearing are all, underneath the politics, arguments about where Australia sits on this curve and how large the behavioural response would be to moving it.
Frequently asked questions
Does this mean tax increases never raise more revenue?
No. It means the revenue gain is typically smaller than static, pre-behavioural forecasts predict, because taxpayers adjust. Whether a specific increase still raises meaningful net revenue depends on where the rate starts and how mobile the affected income is.
What is "income shifting" and why does it matter more than reduced work hours?
It's the practice of restructuring how income is received, into capital gains, deferred compensation, trusts, rather than working less. It matters more because it requires no reduction in actual economic activity, just relocation of the same income into a lower-taxed form, making it a larger and more consistent source of revenue erosion than labour supply effects.
Why does capital income respond more to tax changes than wage income?
Capital owners have far more control over the timing and structure of their income, when to sell an asset, which jurisdiction to hold it in, how to structure ownership, than a wage earner has over their hours, making capital income considerably more tax-elastic.
Does Australia's Treasury already account for this when costing tax policy?
Yes, Treasury's revenue costings incorporate behavioural adjustments, which is an implicit acknowledgment of the effect even though the exact size of the adjustment for any specific proposal remains a genuine point of debate.