An Estonian employer adds 33.8% on top of your salary. A Texas employer adds 7.8%.

By Skyler Bissell · August 27, 2026 · 9 min read

To put €130,000 on an Estonian contract, an employer writes a cheque for about EUR 173,940. The extra is social tax and unemployment insurance, it is compulsory, and it appears on no payslip you will ever be shown. Run the same salary through Texas and the employer's extra is 7.8% of gross.

That gap is why the phrase "tax rate" stops meaning one thing the moment you cross a border. Estonia looks like a low-tax country from a payslip and a heavy one from a payroll. Both readings are arithmetically correct. Which one you reach for decides which country you tell your family is expensive.

Three rates describe one salary

Take one person: a single filer, no kids, no expat regime, on the local equivalent of $150,000. Three different numbers can honestly be called their tax rate.

The headline rate is the top bracket their income touches. It is the number in newspaper coverage and it is the least useful of the three, because almost nobody pays their top bracket on their whole income. It is the first of three numbers our pillar on how countries tax your salary takes apart, and marginal vs effective tax rate covers why it overstates a bill.

The effective rate on contract gross is income tax plus employee payroll contributions, divided by the salary on the contract. This is the number cityparity publishes everywhere else, because it is the one that maps onto a payslip.

The effective rate on employer cost is the same money divided by gross plus the employer's contributions. This is closest to the OECD's tax wedge, and it is the number that makes European systems look heavier and American ones look lighter than the payslip does.

Six cities, all three columns, same person:

Single filer, no kids, no expat regime, no retirement deferral, at each city's local equivalent of $150,000. Employee side computed by cityparity's engine. Employer side from each authority's 2026 schedule with caps applied at this salary. The full 34 rows, sortable, with the components and sources for every jurisdiction, are at employer cost by country.

Look at Tallinn and Copenhagen side by side. On a payslip Estonia is the cheaper country by 16 points. Count the payroll and the two swap places, because Denmark's employer charge is 1.1% of gross and Estonia's is 33.8%. Denmark runs one of the most expensive welfare states in the world and collects almost none of it from employers. It bills the employee through income tax and shows the whole thing on the payslip, which is a design choice about visibility rather than about generosity.

The denominator problem, and why it trips everyone

Here is the trap. A Finnish employer pays a TyEL pension contribution of 17.1 points of gross. Add the other branches and the whole Finnish employer charge is 19.9%. So the two rates for Helsinki should be about 20 points apart, and they are not: the rate on contract is 48.0% and the rate on employer cost is 56.6%, a difference of 8.6 points.

Nothing is missing. The employer contribution enters both the numerator and the denominator. It raises the state's total cut, and it also raises the base you are dividing by, because employer cost is now about 120% of gross. The same money, spread over a bigger base, produces a smaller-looking move.

This is worth understanding before you quote either figure at anyone, because the two rates are not comparable to each other, only to the same rate in another country. Set a French rate on employer cost against a US rate on contract gross and you have manufactured a gap out of nothing but arithmetic. It happens constantly in relocation threads.

The clean rule: compare a column to itself. Every row of the table above uses the same denominator inside each column, so Paris against Austin in the third column is a real comparison and Paris in the third column against Austin in the second is not.

Why caps break the idea of a country rate

Half of these schedules stop charging above a threshold, which means there is no such thing as "the German employer rate". There is only a rate at a salary.

• Amsterdam, Netherlands: five employer premiums, 10.1% of gross in total, and every one of them stops at EUR 79,409. Above that line the marginal employer cost of a raise is zero.
• Berlin, Germany: pension and unemployment cap at EUR 101,400, health and care at EUR 69,750. At a EUR 130,000 salary the marginal employer rate on every statutory branch is already zero.
• Tokyo, Japan: health caps at a JPY 1,390,000 monthly standard remuneration and pension at JPY 650,000, so the average employer rate is 8.4% while the marginal rate at this salary is 1.15%.
• Seoul, South Korea: an average of 7.7% against a marginal 5.97%, because the pension cap binds and the health contribution does not.
• Stockholm, Sweden: 31.4% on the first krona and the same on the last. Belgium, Portugal, Finland, Hungary and Romania are uncapped in the same way.

So a headline like "employers in country X pay 30%" is only meaningful with a salary attached. At a median wage the Dutch employer charge is close to its nominal rate; at a senior engineering salary it is a fraction of it. This is the same effect that makes a marginal bracket a bad summary of an income tax bill, and it runs in the opposite direction: income tax gets heavier as you climb, capped contributions get lighter.

A statutory rate is not what gets collected

Every figure on this page is statutory: what the law says is owed on a standard employment contract. That is the sticker price. Three things move real collections away from it, and none of them is modelled here.

Contract form. A Polish engineer on ryczalt, a German Freiberufler, a Dutch ZZP or an Irish limited company faces a completely different schedule, usually carries both halves of social insurance, and has no employer side to add. Contractor arrangements are how a large share of engineers in central and eastern Europe are paid, so any statement about "what Poland charges" that assumes an employment contract is describing one of two common realities. cityparity prices employment contracts only, and that is a real hole rather than a rounding error.

Regime elections. Spain's Beckham law, the Dutch 30% ruling, Italy's impatriati regime and Portugal's IFICI all cut a qualifying newcomer's bill sharply, and the calculator models them. They are switched off on this page on purpose, because they expire and the salary you live on afterwards is the one worth comparing. Whether one survives contact with a real offer is a question about your own dates rather than about the rate table.

Enforcement and structuring. Effective collection differs from statute in every country, in both directions, and the gap is not something a rate table can show. It is a fair objection to any figure on this page. The defensible claim is narrower: this is what the law charges an employer for a standard contract at this salary, from the authority's own 2026 tables.

The charge that lands in your own account

One rule does more work than the rest, and it is the one people write in to argue about. A mandatory charge that goes to the treasury and a mandatory charge that goes into an account with your name on it are different things, so the employer totals here exclude the second kind and name it in the row instead.

That decision is why four countries sit near zero on this page while their employers pay real money. Singapore's employer CPF is 17% of ordinary wages up to a SGD 8,000 monthly ceiling, which is about 8.5% of this salary, and the treasury's own claim is the Skills Development Levy at 0.1% of gross. An Australian employer pays 12% superannuation. A Hong Kong employer pays MPF. An Emirati employer accrues end-of-service gratuity. In each case the money is yours later, so counting it as tax would misstate what the state takes, and ignoring it entirely would misstate what employment costs. Naming it is the only honest option.

The same logic keeps Italy's TFR and Austria's Abfertigung Neu out of those two rows, and it puts France's AGIRC-ARRCO in, since a points-based pay-as-you-go pension is a transfer rather than a personal balance.

What the rate does not tell you is what the money buys

Belgium is the only row in the set above 60% on employer cost. Read alone, that number is an argument for leaving. Read next to a childcare bill it is something else, because a Brussels household is not paying $2,000 a month for a daycare place and a US household frequently is.

That is the comparison the tax rate cannot make on its own, and it is the reason cityparity exists. The rankings that hold the other half are childcare costs by country and parental leave by country. The post that runs the arithmetic across the Atlantic on one household is the true tax burden once healthcare is counted, and it is the closest thing here to a complete answer: the Seattle to Berlin gap narrows by about a quarter once US health premiums join the tax line.

A fuller version of that answer is in progress: an index that prices what a household actually pays for childcare, healthcare and leave in each country, and sets it against the tax rate on the same salary. It publishes soon and it is the number this page keeps pointing at without being able to name.

Who really pays the employer side

The question underneath all of this is incidence, and it deserves a straight answer about what is known and what is not.

The mainstream finding in labour economics is that employees bear most of the employer payroll charge over time, because employers set gross offers with the total cost in view. If that is right, the third column of the table above is closer to your true burden than the second one. It is also a claim about long-run averages across a labour market, and it says nothing reliable about one senior engineer negotiating one offer in one city.

What you can do with the number is narrower and more useful. If you are being recruited into Estonia or France, the employer is spending far more than your gross, and knowing the total is leverage in a conversation about what the package is worth. Our guide to salary negotiation abroad uses the equivalence figure the same way. If you are comparing two offers in two countries, use the second column, since that is the one that maps onto the money arriving in your account, and then let the safety net decide the rest.

Where these figures come from

The employee side is computed rather than looked up. Each salary runs through its own country's brackets, thresholds and employee contributions in the same engine behind the calculator, on the household stated under the table. The methodology page sets out how, and what it deliberately leaves out.

The employer side is read from each authority's own 2026 tables: URSSAF for France, HMRC for the UK, the Estonian Tax and Customs Board for Estonia, and IRS Publication 15 plus the Texas Workforce Commission for Austin. All 31 authorities are listed row by row on the ranking page, with the components each total is built from and the confidence attached to it. One row carries an estimate rather than a statutory rate, Hong Kong, and it says so.

See all 34 rows, sortable, and move the salary and household yourself →

Or put your own two cities and your real household in: New York against Berlin, or any pair in the calculator. The tax line is one row of what comes back.

FAQ

Which country has the highest employer social security contributions?

France, on this set of 31 countries at a $150,000-equivalent salary. A Paris employer pays 41.7% of gross on top of the salary, once the URSSAF branches, AGIRC-ARRCO complementary pension and the two payroll taxes are added. Estonia is second at 33.8%, on a single 33% social tax plus 0.8% unemployment insurance. Sweden is third at 31.4%, uncapped. At the other end, Denmark charges 1.1% because it funds its welfare state through income tax instead, and Romania charges 2.3% because it moved almost the whole burden onto the employee in 2018.

Do employer payroll taxes come out of my salary?

Not from the salary on your contract, and probably from the salary you were offered. Most labour economists find that employees bear the bulk of employer payroll charges over time, through a gross offer set lower than it would otherwise be. That is a claim about incidence and it is not something a tax engine can settle, so cityparity does not model it. What you can check is the size of the thing being argued about: in Paris it is 41.7% of gross and in Austin it is 7.8%, which is a spread of more than 30 points of salary.

Why do employer contributions stop above a certain salary?

Because most countries cap the assessment base for contributions that fund a capped benefit. A Dutch employer pays five separate premiums and every one stops at EUR 79,409, so the marginal employer cost on a euro above that is zero. Germany caps pension and unemployment at EUR 101,400 and health and care at EUR 69,750, which leaves a marginal employer rate of zero on every statutory branch at a EUR 130,000 salary. Japan caps health at a JPY 1,390,000 monthly standard remuneration and pension at JPY 650,000, giving a marginal rate of 1.15%. Uncapped systems behave the other way: Sweden, Belgium, Portugal, Finland and Hungary charge the same rate on the last euro as on the first.

Does the employer contribution count as tax?

The OECD counts it as tax and calls the combined total the tax wedge. Statistical agencies count it as a labour cost. Both readings are used in public argument, which is why the same country can appear at two very different rates depending on which number the writer picked. cityparity publishes both and labels them, since the honest position is that there is no single true rate: there is a rate on contract gross and a rate on employer cost, and they answer different questions.

Figures come from cityparity's per-city engine and from the statutory schedules listed on the ranking page, each with its issuing authority and the date it was verified. Employer schedules are 2026 rates and most of these jurisdictions reset them in January. Treat any single number as a strong estimate and run your own. See the methodology.