Take-home comparison methodology (2026)

What is compared, and what is not

What is compared is deductions: the taxes and social contributions withheld from an employee's gross salary. What is not compared is gross pay levels, living costs, and what the deductions buy. Those three differ between any two countries more than the deduction rate does, so a lower rate does not mean more money at the end of the month, still less a better life.

Concretely: German contributions buy largely free healthcare at the point of use, the US figures exclude state income tax, which ranges from nothing to over 13%, Australian superannuation is paid on top of salary rather than withheld from it, and Swiss health insurance is not in the deductions at all because it is bought privately per person. None of that compresses into a percentage, which is why every page states both countries' assumptions in full.

The engines

Each country has its own engine implementing the statutory rules for its tax year, with test cases checked against the tax authority's own published examples. The comparison pages call those same engines. There is no second, simplified calculation for comparison purposes and no copied estimates, so a figure on a comparison page matches the one the country calculator gives for the same input.

Standing assumptions, repeated on every page that uses them:

What an individual engine deliberately leaves out, such as the Danish skatteloft or the Belgian werkbonus, is recorded on that country's own methodology page, listed on the methodology index.

Three salary scales

A comparison at an equal nominal amount is only meaningful between currencies of a similar size. One krona is worth about 9 cents and one yen well under a cent, so holding 60,000 constant would put Sweden, Norway, Denmark and Japan at the top of any ranking on near-zero deduction rates, because 60,000 kronor is about €5,469 and 60,000 yen about €337 a year. That would be an artefact of currency size, not a result. Hence three scales:

The Nordic currencies are within a factor of two of each other, so Sweden against Norway or Denmark compares nominally and involves no exchange rate at all. Those same countries against the euro and sterling countries have no such option, so those pages convert. Japan has no partner of its own size at any scale, so every Japan page converts.

The five points above are the rows shown on each pair page. The per-salary pages use a wider grid of 9 round amounts, since round numbers are what people search for: 30,000, 40,000, 50,000, 60,000, 70,000, 80,000, 100,000, 120,000, 150,000 on the standard scale, 300,000, 400,000, 450,000, 500,000, 600,000, 700,000, 800,000, 1,000,000, 1,200,000 on the Nordic one, and 30,000, 40,000, 45,000, 50,000, 60,000, 70,000, 80,000, 100,000, 120,000 euro when converting.

Exchange rates

Conversion uses the European Central Bank's daily euro reference rates, which are free, need no key, and are the reference most European institutions quote. The set is pinned and committed alongside the source, currently 2026-11-04, and refreshed deliberately by a script rather than fetched during a build. That is a considered trade: a rate fetched at build time would make published figures unreproducible and would let a network blip change them silently. The cost is that the euro figures drift from the market as time passes, which is why the date appears on every converted page.

One point is easy to miss and stated on the pages themselves: the rate does move the deduction percentages, just not directly. The tax rules are applied in local currency and never see the rate. But the rate decides which local salary each row lands on, and every system here is progressive, so a weaker currency pushes the converted rows further up the scale and the percentage rises with them.

Denmark, the one pegged currency

One of these conversions is sturdier than the rest. Denmark holds the krone against the euro under ERM II at a central rate of 7.46038 per euro, inside a multilaterally agreed band of 2.25%, which allows it to move only between 7.29252 and 7.62824; both central banks are obliged to intervene at the edges. The rate pinned here sits 0.18% from the central rate. Danish euro figures therefore cannot be invalidated by a currency swing in the way a floating conversion can. The krona, the Norwegian krone and the yen carry no such commitment, and for them the pinned figure is simply the market rate of that day. Checked against Danmarks Nationalbank's own publications and the European Central Bank's rate table, which agree.

The equivalent gross

The more useful question is usually not what the same gross leaves, but what you would have to be paid on the other side to be no worse off. That figure is found by running the destination engine backwards: bisecting until its net matches the target. The whole deduction structure is therefore in the answer, allowances, bands and contribution ceilings included, rather than an average rate applied to a difference.

It appears on 21 of the 39 pairs, namely those where both net figures are in one unit: the converted pairs, in euros, and pairs of two euro countries. On the other 18 it is deliberately withheld. A nominal comparison across two currencies holds the same number constant on purpose, so matching one net against the other there would match a number and not an amount of money. Those pages say why the question cannot be answered rather than printing a figure that looks credible and is meaningless.

Effective and marginal rates

The effective deduction rate is gross minus net, over gross. The marginal rate is the share of the next slice of gross pay that does not reach take-home, measured by raising gross by a round step and looking at how net responds. The step follows the currency so the wording stays readable: 1,000 for the euro and similar currencies, 10,000 for the Nordic ones and 100,000 yen. These are not equal amounts of money, 100,000 yen being about 561 euro, but each is a natural unit in its own currency, and the choice barely affects the result because the marginal rate is near constant over so small a step, except right at a threshold.

In a progressive system the marginal rate always exceeds the effective one, because allowances and lower bands hold the average down while the last slice of pay has already passed them. That gap is why a raise rarely feels like its headline size, and it is asserted as a test for every country at a mid salary.

Limits and known gaps

• Figures are for a single person with no children and no other income. Joint assessment, child allowances and secondary income move the result substantially; the country calculators expose those, these pages do not.
• Regional rates are fixed to one value, such as Zurich for Switzerland, North Rhine Westphalia for Germany, Ontario for Canada, and the national average municipal rate in Sweden and Denmark. An actual municipality may differ.
• Employer contributions are excluded, since they are not withheld from gross pay. That understates the total cost of employment in countries where much of the burden sits with the employer.
• Individual statutory rules are omitted where they could not be implemented reliably. Each is named on the relevant country methodology page and in the calculator's own notes.
• Exchange rates are pinned to a date and do not update on their own.
• German versions exist for the 15 pairs involving Germany, Austria or Switzerland, at Gehaltsvergleich nach Ländern. The remaining pairs are English only.

Updating and sources

Tax rules are taken per tax year from the responsible authority's own publications rather than third-party summaries, and the per-country methodology pages name the documents used. Exchange rates are pulled from the European Central Bank by a script and committed, so every publication stays reproducible. Net figures on this page and every comparison page are recomputed from the engines on each build.

Frequently asked questions

Last updated 2026-11-05. Back to the take-home comparison or the methodology index.