What comes with the job, and who tells you.

Most job postings tell you something about the offer beyond the pay, but not all of them, and not the same things. We took 20,575 US postings from employer career sites, dated August 31 to September 27, 2026, and looked for four benefits a job seeker would want to know about before applying: dental cover, a retirement plan, paid time off and parental leave. 56.7% name at least one. Where a state’s law asks the posting to describe benefits, 70.8% do; where the law asks for pay only, 57.9%; where there is no posting law, 52.1%.

Issue 03 measured who tells you the pay. This issue measures the rest of the offer: how often each benefit is named, how that changes with the law where the job is, whether the same employer writes differently in different states, how benefits travel with pay, and which kinds of work say least. No employer is named. Section 08 explains why this issue counts career-site postings only, and the correction to Issue 03 that this led to.

Just over half name a core benefit

Most postings carry some kind of benefits list: 72.4% list at least one item. But the lists include things that are not benefits in the usual sense, like “competitive salary” or “training”, so this issue counts four specific benefits instead. A retirement plan is named in 51.4% of postings, paid time off in 44.7%, dental cover in 44.7% and parental leave in 14.6%. At least one of the four appears in 56.7%.

The four tend to come together. 35.4% of postings name dental cover, a retirement plan and paid time off all at once, and 10.2% name all four. When a posting has a benefits list it holds 9.2 items on average. At the other end, 27.6% of postings list nothing at all.

The list comes from our data vendor, which reads the full posting. We checked it against the posting text on the 11,682 postings (56.8%) whose whole text we hold. When the list says dental cover, the text says so 98.8% of the time, and when the text names dental cover, the list carries it 97.7% of the time. For a retirement plan the two figures are 95.7% and 99.3%, for paid time off 97.7% and 95.7%, and for parental leave 93.2% and 96.1%. An empty list means the posting is silent: of the 3,689 whole postings with no list, 2.6% name any of the four anywhere in their text.

Where the law asks for benefits

Pay-transparency laws differ in what they ask of a posting. Six states require a pay range and a general description of the benefits that come with the job: Colorado, Washington, Illinois, Minnesota, New Jersey and Maryland. Illinois lets a posting meet that requirement with a link to a benefits page, so a posting there can comply and still name nothing. Eight jurisdictions require the pay range only: California, New York, Hawaii, the District of Columbia, Vermont, Massachusetts, and since this summer Virginia and Maine. Every other state has no posting law. All of these laws were in force for the whole window. The city of Cleveland, whose ordinance asks for a salary range and counts benefits as part of salary, fits neither group; its 41 postings count in the totals only.

Postings in the six pay-and-benefits states name a core benefit 70.8% of the time (2,826 postings). In the pay-only jurisdictions it is 57.9% (5,205), and in states with no posting law 52.1% (11,450). Postings with no state recorded are at 62.3% (1,053) and are not pooled into any group.

Two checks. The kinds of work posted differ from state to state, so we gave every group the same mix of role families as the whole universe: the three shares become 70.9%, 58.1% and 52.4%. Some employers post far more than others, so we counted no employer more than ten times in any group: 72.0%, 59.6% and 54.6%. The order does not change.

Nor is this simply a matter of more open employers. The pay-only jurisdictions state pay more often than the pay-and-benefits states, 77.7% against 74.2%, and name benefits less often. Each of the four benefits follows the same order across the three groups: a retirement plan is named in 65.5%, 53.1% and 46.8% of postings, paid time off in 55.9%, 43.9% and 42.1%, dental cover in 55.8%, 44.8% and 41.8%, and parental leave in 17.9%, 17.4% and 11.6%.

By state

Twenty-two states have 300 or more postings in the window. The four highest are all pay-and-benefits states: Washington at 81.6%, Minnesota at 74.6%, Illinois at 72.1% and Colorado at 72.0%.

Below the top four the line blurs. Arizona, which has no posting law, is fifth at 66.6%, above two pay-and-benefits states, Maryland (64.8%) and New Jersey (61.3%). The two largest pay-only states sit among the no-law states: California at 56.9% and New York at 56.5%, level with North Carolina (56.5%), Michigan (56.2%) and Ohio (55.4%). Florida is lowest at 46.6%. The difference between the groups in section 02 is real, but it is not a clean line between states. In no state on the chart is a single employer more than 7.2% of postings.

The same employer, two kinds of state

The state pattern could be an employer pattern in disguise: perhaps the employers that post in Washington or Illinois are the kind that list benefits everywhere. Many employers post in more than one kind of state, so we can check. 94 employers have at least three postings in a pay-and-benefits state and at least three in a state with no posting law. Their postings in the pay-and-benefits states name a core benefit 66.2% of the time (710 postings), and their postings in no-law states 45.4% (2,370).

Employer by employer, benefits move less than pay. 44 of the 94 state pay at least five points more often in their pay-and-benefits postings, 42 are within five points either way, and eight go the other way. For benefits, 24 name a core benefit at least five points more often on the pay-and-benefits side, 61 are within five points, and nine go the other way. The pooled figures count postings, not employers, so the minority of employers that write differently on each side can move them a long way.

Between the two kinds of law state, the same check shows little difference. 68 employers have at least three postings in a pay-and-benefits state and three in a pay-only jurisdiction: 62.2% of their pay-and-benefits postings name a core benefit (579) and 59.0% of their pay-only postings (910), and 44 of the 68 are within five points. So for most of these employers, the gap is not them writing differently in different places; this issue does not test what it is instead. A stricter threshold of five postings on each side gives 45 employers at 62.5% and 41.7% against no-law states, and 30 employers at 63.1% and 63.2% against pay-only jurisdictions.

Benefits travel with pay

Postings that state pay name a core benefit 70.0% of the time (10,796 postings); postings that do not, 42.1% (9,779). That holds in every law group and in every one of the 23 role families with 300 or more postings.

The law groups differ mostly among the postings that state pay. Where the law asks for pay and benefits, 80.5% of postings that state pay name a core benefit; where it asks for pay only, 63.9%; with no posting law, 69.9%. Among postings that state no pay the three groups sit closer together, at 43.0%, 37.3% and 42.7%. One possible reading is that where the law asks for both, the block of text that carries the pay range carries the benefits too; this issue does not test it.

The postings that tell you least name neither a pay figure nor any of the four benefits: 27.5% of all postings. In states with no posting law it is 37.5%; where the law asks for pay and benefits, 14.7%; where it asks for pay only, 14.0%.

By kind of work

Real estate and property postings name a core benefit most often, at 71.4%, followed by therapy and behavioral health (70.8%), registered nursing (65.7%) and skilled trades and facilities (65.4%). At the bottom are banking, lending and insurance (40.7%), food service and hospitality (40.9%), customer service and support (43.3%) and general and location management (45.0%).

Parental leave varies widely. It is named in 22.5% of software engineering postings, 22.2% of operations and supply chain postings and 21.2% of data and analytics postings, against 3.9% of food service and hospitality, 6.3% of general and location management and 6.9% of customer service postings.

In every one of the 23 families, postings that state pay name a core benefit more often than postings that do not. Where both groups have 300 or more postings, the gap is widest in banking, lending and insurance (69.1% against 21.4%), education and childcare (74.3% against 28.8%) and customer service and support (68.9% against 26.3%), and narrowest in accounting, audit and tax (66.5% against 52.6%).

What to do with this

• Treat a silent posting as a question. 43.3% of postings name none of the four benefits, and when the list is empty the text almost never names one. Ask about health cover, retirement, paid time off and parental leave early, before you put hours into the process.
• Expect more where the law asks for it. In Washington, Minnesota, Illinois and Colorado, 72.0% to 81.6% of postings name a core benefit. The laws exempt some smaller employers, which we cannot see, so a silent posting there is not necessarily unlawful, but it is less common.
• Read pay and benefits together. 70.0% of postings that state pay name a core benefit, against 42.1% of postings that do not. A posting with neither tells you least about the offer: 27.5% of all postings, and 37.5% in states with no posting law.
• Don’t expect another posting to fill the gap. 61 of 94 employers that post in both a pay-and-benefits state and a no-law state name benefits at about the same rate in both, within five percentage points.
• Ask about parental leave directly. It is named in 14.6% of postings. A posting that does not mention it tells you nothing either way.

How we counted

The short version. The long version, with every definition, lives on one page and every report links to it: About Longbow Research.

• One window, career sites only.Postings dated August 31 to September 27, 2026, published on employers’ own career sites and hiring systems: 20,575 postings from 8,023 employers. Our feed carries no benefits list for job-board postings (1,922 in the window, none with a list), so they are left out rather than counted as silent.
• Frozen, so it can be checked. The postings behind this issue were copied out of our feed on September 30, 2026, one row per posting with every field the queries read, and the copy is kept in the Longbow repository. Every figure is computed from that copy, so it can be re-run and comes out the same long after the feed has deleted the postings.
• The list is read from the full posting.Our own stored copy of a posting is cut at 5,000 characters; the vendor’s benefits list is read from the whole posting. Where we hold the whole text, the list and the text agree between 93.2% and 99.3% of the time, in both directions, for all four benefits.
• Four benefits, not medical cover.Postings word medical cover in too many ways (“medical, dental and vision”, “health benefits”, “comprehensive benefits”) for the same check against the text, so it is not one of the four.
• Law means law on the posting date.The methodology lists each law with its citation. Most exempt small employers, and the feed carries no employer-size value for these postings, so “law in force” means in force in the jurisdiction, not necessarily binding on that employer.
• No group under 300 postings. That is why nine families, not 23, appear on the pay-gap chart, and why some states are missing from the state chart.
• No names, no pay levels, no causes. No employer is named and no pay figure is published. The law comparison describes where benefits are named; it does not show that a law made the difference.
• A correction to Issue 03. Issue 03’s benefits figures counted job-board postings, which carry no list, as listing no benefits. Measured again without them on its fall window, its finding that benefits travel with pay holds; its finding that a posting with no pay in a law state is the least likely to list benefits does not. The correction is dated on Issue 03’s page.

Coming next. Issue 05, What Employers Actually List: the skills named in postings, by role family. Each issue goes out as an edition of The Longbow Index on LinkedIn the day it publishes.