Who Actually Pays

A payroll tax is split between employer and employee. In many countries the legislation says so explicitly: half from each.

The economics says the split is essentially irrelevant. What determines who bears a tax is not the legal obligation but relative elasticity, which side of the market is less able to change its behaviour in response.

The logic is straightforward once you see it. If a tax raises the cost of employing someone, and workers cannot easily withdraw from the labour market, the employer reduces wages until the after-tax cost is back where it was. If workers could easily stop working, they would, and the employer would have to absorb more. Labour supply is fairly inelastic in most settings, which is why studies generally find that payroll taxes fall substantially on wages regardless of the legislated split.

The same logic runs everywhere. A tax on a good with very inelastic demand (tobacco, fuel, insulin) is paid by consumers, because quantity barely responds and the seller can raise the price by nearly the full amount. A tax on a good with elastic demand falls more on producers, because raising the price loses too much volume.

The limiting case is the most interesting, and it explains a policy that economists across the political spectrum have endorsed for a century and a half.

The supply of land is perfectly inelastic. Nobody can make more of it in response to a higher return, and nobody withdraws any in response to a lower one. So a tax on land value cannot be passed to anybody: it falls entirely on the owner, and what happens is that the price of the land drops by the capitalised value of the tax.

Henry George built a political movement on this in the 1880s, and the economic case has never been refuted: a land value tax raises revenue without discouraging any productive activity, because the thing being taxed cannot be produced or destroyed. Its problems are practical and political, separating land value from the value of the buildings on it is genuinely difficult, and every existing owner has a large and immediate reason to object.

The honest case in the other direction is corporation tax, where the question is unresolved and the disagreement is technical rather than ideological.

A company is a legal fiction; the burden must land on people. Candidates are shareholders through lower returns, workers through lower wages, and consumers through higher prices. Estimates of the share borne by workers range from around 20% to more than 50%, depending on assumptions about how mobile capital is internationally, how competitive labour markets are, and the time horizon considered.

Anybody who tells you confidently that corporation tax is paid entirely by capital, or entirely by workers, is reporting a political position. The range in the literature is genuinely that wide.

Two further points make the overall picture harder than the headlines.

First, progressivity is a property of a system rather than of a tax. A tax is progressive if it takes a rising share of income as income rises, and regressive if the share falls. Value added tax is nominally flat (the same rate for everyone) and regressive in effect, because poorer households spend a larger fraction of their income and so pay a larger fraction of it in a tax on spending.

This means an income tax can be steeply progressive while the system as a whole, including consumption taxes, payroll taxes and local taxes, is close to flat. Several rich countries are in roughly that position, and looking at the income tax schedule alone will tell you the opposite.

Second, the announcement effect is not the incidence. When a city taxes landlords and the landlords announce they will pass it to tenants, the announcement tells you what landlords would like. Whether they can depends on whether rents were already at the level the market will bear. If housing supply is fixed in the short run and demand is strong, rents were already as high as tenants could pay, and a tax creates no new willingness to pay; the burden stays with the owner.

Which is a general rule worth carrying: whenever anybody announces who will pay for something, the question to ask is what would have to be true about behaviour for that to happen.