A salary range in a job posting is the pay a company says it expects to offer for one opening, on the day it posts. What it actually pays the person hired can sit anywhere in that range, or well away from it once variable pay is counted. This post sets out what statistics offices and central banks have measured, five reasons the two part, and a test for each to run on your own offers.
How close are salary ranges in job postings to what companies actually pay?
We found no official study that follows a posted range to the pay of the person hired into that post. The nearest official measures compare postings with workers: in lower-paid US occupations, posted wages averaged about 40 percent above employed workers' wages in a 2026 staff report, while Canada's offered wage ran below the average hourly wage in 2024.
We searched the published research and statistics of the Federal Reserve Banks, the US Bureau of Labor Statistics, the Bank of England, the European Central Bank, Statistics Canada and the Office for National Statistics. Each comparison we found sets postings beside the pay of people already in work, not the offer made on the same opening.
| Who speaks | Date | Population | What is compared | Finding |
|---|---|---|---|---|
| Batra, Michaud and Mongey, in a Federal Reserve Bank of Minneapolis staff report; the authors' finding | Staff Report 690, 28 September 2026 (earlier as Institute Working Paper 83, 21 December 2023) | US online job posts | Posted wages against workers' wages in the Current Population Survey and the occupational wage survey | 14 percent of posts state any pay; posted averages around 40 percent higher, especially at the bottom of the distribution, and 80 percent in low-wage salaried positions |
| Statistics Canada | Third quarter of 2024, released 16 December 2024 | Vacancies in the Job Vacancy and Wage Survey; all employees in the Labour Force Survey | Average offered hourly wage against average hourly wage | 27.55 against 35.24 Canadian dollars |
| Statistics Canada | Second quarter of 2026, released 15 September 2026 | The same surveys | Year-on-year growth | Offered wage up 2.0 percent to 28.55; all employees up 3.6 percent |
| Office for National Statistics | Asking salaries released 18 July 2025, January 2017 to May 2025 | UK online job adverts | Published apart from the earnings survey | We found no ONS comparison of the two |
The two comparisons point in opposite directions, and they are built differently. The US study averages the pay stated in posts that state any. Statistics Canada's offered wage is the lowest value of a range, as its survey guide (2024 edition) asks employers to report it, and leaves out overtime, tips, commissions and bonuses, while its Labour Force Survey wage includes tips and commissions. Statistics Canada notes that the offered wage may differ from the wage actually paid once the position is filled. The gap, and even its sign, depend on the figure taken from a range, the pay elements counted and the workers compared.
What an advertised salary range is
An advertised range is a statement of intent: the pay an employer expects to offer for one posting, on the date of posting. It records no payment. In New York City the Commission on Human Rights states that since 1 November 2022 the city's Human Rights Law has required a good faith pay range in every job advertisement, and the New York State Department of Labor describes a range as the minimum and maximum believed to be accurate at the time of posting; for any decision, read the official text for your own jurisdiction.
Read as data, a range therefore has a date, a scope (base only, or base with variable pay, as the employer chose to state) and a population: openings advertised, not jobs already filled.
Offers, ranges and pay transparency research
Do employers make actual pay offers within the posted salary range, at its midpoint?
We found no official research that measures where actual offers land inside posted ranges, so the midpoint is an analyst's convention, not a finding. The width matters: in US online posts that list a range, the Minneapolis Fed authors found ranges spanning on average 28 percent of the midpoint, so either end sits about 14 percent from it.
A midpoint benchmark and a low-end benchmark built from the same postings differ by half the range's width before any offer is made. The test is the position in range of each offer: the offer less the posted minimum, divided by the maximum less the minimum, where 0.5 is the midpoint. Compute it against the range each hire answered, and report the median and spread with the count of hires.
What does pay transparency research show about employers' actual offers against salary ranges in job postings?
The official research we found on pay transparency measures whether postings state pay, not where offers land. In a New York Fed blog post, Richard Audoly and Roshie Xing found that, as of January 2025, about 24 percent of job ads covered by transparency rules in four US jurisdictions gave no salary. We found no official study comparing those posted ranges with offers made.
The post of 2 October 2025, whose views are its authors' own, covers Colorado, California, Washington and New York City, and finds the share of postings with salary information rose by an average of 20 percentage points in the month each rule took effect. So the employers stating pay in a place change on the date a rule starts: compare medians either side of that date before reading a move as a change in pay.
Has a study of a salary transparency law compared the posted pay range with the actual offer?
We found no such study from a statistics office, a government department or a central bank. The official work we read measures compliance and coverage: whether ads covered by a law state pay at all. To measure the gap for your own roles, set each posted range beside the offer you made for it.
Five reasons advertised and actual pay part, and a test for each
| Reason | What it does to a benchmark | Test on your own data |
|---|---|---|
| The offer lands somewhere in the range | A midpoint benchmark assumes offers centre on the midpoint | Position in range of each offer; report the median and the spread |
| Ranges are wide | A range spanning 28 percent of its midpoint leaves 14 percent either side | Width as a share of the midpoint for each posting; set aside ranges wider than your own policy before taking midpoints |
| Variable pay | Advertised on-target earnings are a target; attainment decides what is paid | Realised total pay divided by the stated on-target earnings, by attainment band |
| Only some postings state pay | Postings that state pay are a selected set of employers and occupations | Share of postings stating pay by role and employer; recompute the median without the largest employers |
| Only open roles are seen | The mix of open roles differs from the mix of filled jobs, and moves the average | Compare the mix of roles and seniority in the benchmark sample with your own headcount, then reweight |
Variable pay matters most for sales roles. On-target earnings are base pay plus the variable pay a seller earns at quota, so a seller below quota takes home less than the posting states. Official statistics differ here too: the US occupational wage survey counts commissions and production bonuses as wages, while the Canadian offered wage leaves them out.
Selection is the second large effect. The Minneapolis Fed authors report that 75 percent of all posts come from firms that state pay in less than 5 percent of their posts, and that postings for high-wage occupations and firms use ranges more often, and wider ones, when they state pay.
The mix of open roles moves averages on its own. Statistics Canada attributed part of the rise in offered wages in the third quarter of 2024 to a shift of vacancies from lower-paid to higher-paid occupations: that part of the rise owes nothing to what any one job pays.
Worked example: one posted range against five offers
The example is illustrative. Acme Robotics is an invented company and every figure is invented to show the arithmetic.
Assumptions:
- Acme Robotics advertises an account executive role with a base range of 80,000 to 100,000 US dollars a year and on-target earnings of 160,000.
- Five people are hired from the posting, at base offers of 82,000, 85,000, 88,000, 90,000 and 95,000.
- Each plan carries a variable target of 70,000 on top of the base.
- First-year attainment is 85 percent of quota for every hire, paid in a straight line with no accelerators.
The working:
- The midpoint of the range is (80,000 + 100,000) / 2 = 90,000. The width is 20,000, which is 22.2 percent of the midpoint.
- Variable pay at 85 percent attainment is 0.85 × 70,000 = 59,500 for each hire.
- Realised pay is the base offer plus 59,500.
| Hire | Base offer | Position in range | Variable paid | Realised pay |
|---|---|---|---|---|
| 1 | 82,000 | 0.10 | 59,500 | 141,500 |
| 2 | 85,000 | 0.25 | 59,500 | 144,500 |
| 3 | 88,000 | 0.40 | 59,500 | 147,500 |
| 4 | 90,000 | 0.50 | 59,500 | 149,500 |
| 5 | 95,000 | 0.75 | 59,500 | 154,500 |
The median base offer is 88,000, which is 2,000 or 2.2 percent below the posted midpoint, and the median position in range is 0.40. The median realised pay is 147,500, which is 12,500 or 7.8 percent below the advertised on-target earnings of 160,000.
The midpoint came within 2.2 percent of Acme's median base offer; the larger gap came from attainment, settled after the posting closes. Advertised on-target earnings describe targets, so set your own attainment beside them before comparing them with what your team takes home.
The limits of advertised pay as a compensation benchmark
Advertised pay measures what employers offer for open roles that state pay, at the date of posting. As a compensation benchmark it is limited by where offers land in the range, by range width, by variable pay that depends on attainment, by which employers state pay, and by the mix of roles that happen to be open.
What it measures well is the market's current offer, dated to the week, and its direction. On the ECB Blog of 23 May 2024, four authors whose views are their own read a wage tracker built from job postings, developed with the Central Bank of Ireland, beside the ECB's tracker of negotiated wages, as another indicator that euro area wage pressures were moderating.
Problems with using job postings salary data for compensation benchmarking
Four problems recur: the postings that state pay are a selected share of all postings, a range must be reduced to one figure, hourly, monthly and foreign-currency pay must be put on one scale, and the posting date can precede the offer by weeks. Each calls for a rule you set, apply to every source and report.
Choose the low end, the midpoint or the high end and keep to it. State the yearly hours behind any hourly conversion, and the exchange rate and its date. Date each range by when the posting was first seen.
What are the limitations of advertised salary ranges in job postings as a compensation benchmark in a pay transparency labour market?
Transparency rules raise the share of postings that state pay, which widens a benchmark's coverage, but they can change who is in it at the date each rule starts, and a posted range is not the pay agreed. Read a benchmark in such a market by place and date, and check for level shifts when a rule takes effect.
What official wage statistics measure instead
Official wage statistics measure pay received, by people in work or newly hired, and the offers they record are dated and defined. Set an advertised figure beside the one that answers your question.
| Statistic | Publisher | What it measures | Beside advertised pay |
|---|---|---|---|
| Occupational Employment and Wage Statistics | US Bureau of Labor Statistics | Wages of employed workers by occupation, from a survey of establishments, commissions included | Pay of people in post |
| Quarterly Workforce Indicators | US Census Bureau | Average monthly earnings of new hires who reach full-quarter employment, from unemployment insurance wage records | Pay of the newly hired |
| Annual Survey of Hours and Earnings | Office for National Statistics | Pay of a 1 percent sample of employee jobs drawn from PAYE records, each April, reported by employers | Pay received in the UK |
| Job Vacancy and Wage Survey | Statistics Canada | Hourly wage offered for vacancies, the lowest value of a range, without overtime, tips, commissions or bonuses | Offers, read at the low end |
Compare the same occupation and place on both sides, and track the ratio of advertised to official pay over time rather than the level: the ratio absorbs differences of definition that stay constant.
Pay as the posting states it, in Fokals data
Fokals delivers the pay employers state in postings on their own job boards and careers pages, as why we read job postings at the source explains. Every posting in Job Postings that states pay carries it as written, with minimum, maximum, currency and period, the same pay on one yearly US-dollar scale at a reference exchange rate, and the dates it was first seen, last seen and closed. For a posting in a sales function, base pay and on-target earnings are taken where the posting names them apart, with the base share of the plan, whether commission is uncapped, any quota and the ramp period.
Sales Pay Benchmarks turn those statements into weekly quartiles of advertised base pay and on-target earnings by sales role and country, each group published with the number of postings behind it and written once after the week closes. Market Series carry weekly medians of advertised pay, including sales base and on-target medians, by country, industry and company size band. Job Postings and Sales Pay Benchmarks belong to the hiring product, Market Series to the market series product, and the data dictionary defines every field.
Each of the five tests has the fields it needs. The minimum and maximum as written give the width of every range and leave the reduction rule to you. Base and on-target earnings stated apart separate the base gap from the attainment gap. The postings count on every quartile, and the company on every posting, show how many employers stand behind a cut. The first-seen date dates each range, and the job function and seniority labels let you reweight to your own headcount. Placing a plan against the quartiles step by step is the subject of benchmarking sales compensation from advertised pay, and the methodology sets out how pay is read from a posting.
Frequently asked questions
Is the midpoint of a posted salary range what a new hire is paid?
Not necessarily. We found no official study that measures where offers land inside posted ranges, so the midpoint is a convention. US online postings that list a range span on average 28 percent of the midpoint, according to a 2026 Federal Reserve Bank of Minneapolis staff report, so an offer at either end sits about 14 percent from it. Measure the position in range of your own offers before you assume the midpoint.
Why do advertised salaries differ from official wage statistics?
They measure different things. Advertised pay is an employer's stated expectation for open roles that state pay, on the posting date. Official statistics such as the US occupational wage survey, the UK earnings survey and the US Quarterly Workforce Indicators measure pay received by people in work or newly hired. Which employers state pay, which figure is taken from a range and which pay elements count all move the gap.
What share of job postings include a salary?
It depends on the place and the date. A 2026 Federal Reserve Bank of Minneapolis staff report found that 14 percent of US online job posts contain any pay information. Transparency rules raise it: a New York Fed blog post found the share with salary information rose by an average of 20 percentage points in the month rules took effect in four US jurisdictions, while about 24 percent of covered ads still gave none in January 2025.
Should sales pay be benchmarked on advertised base or on-target earnings?
On both, read together. Advertised base pay is the fixed salary, paid whatever the seller's results. On-target earnings add the variable pay a seller earns at quota, so realised pay moves with attainment. Compare base with base and on-target earnings with on-target earnings, then set your own attainment beside the second before you compare it with what your team takes home.
Can advertised pay replace a compensation survey?
They answer different questions. Advertised pay shows what competing employers offer today for open roles, dated to the week, which suits hiring and offer decisions. A compensation survey built on payroll data shows what people in post are paid. Use advertised pay for the market's current offer and its direction, and payroll-based figures for pay received.
What this page says about the products it names was checked against their public documentation on 7 October 2026. Product and company names are trademarks of their owners. Fokals is not affiliated with them or endorsed by them.