What does an employee actually cost, over and above the wage?

A small employer asking what an employee costs beyond the wage is asking a single, obvious question, and the published record answers it with several different numbers that are not versions of each other. Set side by side they look like a range. They are not a range. Each was measured over a different denominator, for a different group of people, and in three cases they are not measurements at all but rates written into law. The table above puts each figure next to what it actually counted, which is the only way to see that.

The largest single confusion is the denominator. The survey figures are shares of total compensation: the employer's whole hourly cost, with wages sitting in the bottom of the fraction alongside benefits. A great many calculators take a figure of that shape and multiply it by a salary, which is a different fraction with a different bottom. The result understates the load, consistently, and nothing on the publisher's page warns anybody about it, because the publisher never proposed that use.

The second confusion is population, and it is easy enough to make that this project made it in its own working notes before catching it. The figure most often quoted as the employer benefit share describes private industry only. The figure covering all civilian workers — private industry plus state and local government — is a different and higher number, and it appears in a table rather than in any sentence of the release. Government employers pay materially more per hour worked, so which of the two you use changes the answer, and almost nobody says which one they used.

The third problem is the one that matters most to the reader this page is for, and it is the one the page cannot fix. The survey does publish a small-employer cut, and the cut is by establishment — a single physical workplace — rather than by firm. A branch office of a national chain with forty people in it sits in that band next to a genuinely independent business with six. Nothing published says how much of the band is which. So the small-employer figure answers a question about workplaces, and the question a small business owner is asking is about a company. That gap is recorded above as an open decision rather than papered over, because there is no honest way to close it from the documents.

The statutory contributions are a different kind of object again. They are not survey estimates; they are rates set in law and published by the authority that collects them, and two of the three stop at a cap. One stops at an annual wage base per employee and one stops after the first few thousand dollars of each employee's pay, which means what an employer actually pays as a fraction of its wage bill depends on how pay is distributed across its own staff. Any tool that adds the three rates together and multiplies a salary by the total is asserting a share while citing a rate, and it is wrong for every employee whose pay crosses a cap.

The useful conclusion is not a multiplier. It is that the benefit share and the statutory floor are two separate quantities with two separate denominators; that only one of them is measured, and only for workplaces rather than for firms; and that the multiplier in general circulation has no published measurement behind it at all. An employer who wants their own number can get it from their own payroll in an afternoon, and it will be more decision-relevant than any figure here.

One further thing is worth knowing before any of these figures is written into a spreadsheet that will be reused. The publisher has announced that it will remove workers' compensation costs from the survey with the December reference period. Workers' compensation sits inside legally required benefits, which is the component this question cares about most, so the published figure will fall without a single employer's costs changing. Figures either side of that boundary are not comparable, and this page will need rewriting rather than updating when it lands.