You are sending applications into a market where postings are up and hires are down, and the gap between the two is now the story. The September 1, 2026 BLS JOLTS release for July put the hires rate at 3.2%, the same average the US ran in 2010 when unemployment sat at 9.6%. If your response is to double your application volume, you are pouring effort into the exact leak the data just described.
What the July 2026 JOLTS release actually said
Job openings held steady while hiring collapsed, and white-collar work took the deepest cut. Openings rose 89,000 to 7.271 million, keeping the openings rate at 4.4%. Hires fell 294,000 to 5.054 million and the hires rate dropped from 3.4% to 3.2%, the weakest hiring month since February. Professional and Business Services alone lost 188,000 hires month over month.
The quits rate stayed pinned at 1.9%, matching the post-pandemic floor. That number is the tell: workers are not confident enough in the market to walk, which means the internal churn that normally opens seats has stopped feeding the outside candidate pool.
| Segment | Figure | Source |
|---|---|---|
| US hires rate, July 2026 | 3.2% | BLS JOLTS |
| US openings rate, July 2026 | 4.4% | BLS JOLTS |
| Openings-to-hires wedge | 1.2 pp | Derived from JOLTS |
| Openings-to-hires ratio | 1.44 openings per hire | Derived from JOLTS |
| Prof. & Business Services MoM hires change | −188,000 | Indeed Hiring Lab |
| Quits rate | 1.9% | BLS JOLTS |
| Hires-per-posting, long-run drop | 0.8 → 0.4 since 2019 | Columbia Law Review |
Matches the 2010 annual average, when unemployment ran at 9.6%.
Why the hires rate matters more than U-3 right now
The hires rate is the true measure of mobility, and it says the market is frozen. Unemployment stays low because employers are not firing, not because they are hiring. A conspicuous lack of both hiring and firing means there is not much labor turnover, and no turnover means no seats to move into.
For an active job seeker, that reframes the search. The headline unemployment print does not describe your experience. The 3.2% hires rate does. Every month a company keeps a posting live without pulling anyone in adds to the pile, and that pile is now the median of what you are applying to.
The 1.44 openings-per-hire ratio, decoded
For every posting that ends in a hire, roughly 0.44 postings will not, and that is before you strip out true ghost jobs. The math is simple: 7.271 million openings divided by 5.054 million hires equals 1.44. The Columbia Law Review found the hires-per-posting ratio has halved since 2019, from 0.8 to 0.4, so the wedge you feel is real and it has a five-year trend behind it.
The mechanism is not conspiracy, it is repurposing. Postings have become a marketing and reconnaissance activity, not a hiring commitment. Employers use them to:
- Build talent pools for a hire they might make next quarter.
- Monitor specialist availability so they know what a replacement would cost.
- Benchmark salary expectations against the current market.
- Signal growth to customers, investors, or their own employees.
None of those goals involve hiring you. A Clarify Capital analysis of over 175,000 US listings found roughly 1 in 7 postings were ghost jobs in 2026. Greenhouse, the ATS, has said 18 to 22% of jobs posted in any given quarter qualify. Resume Genius found 67% of job seekers suspected they had encountered one.
Where the ghost jobs cluster
Marketing and advertising and mid-level generalist software engineering are the two ghost-heavy zones, and they are also two of the largest supply concentrations in the market. Ghost rates track oversupply almost perfectly.
- Marketing and advertising: 87.5% of professionals report encountering phantom listings, the highest across surveyed industries.
- Mid-level generalist software engineering: the profile of "Software Engineer" with a generic description, no salary range, open 90+ days, from a 2,000-person company is the archetype.
- Companies with 1,001 to 5,000 employees: nearly 25% of their postings qualify as ghosts.
- Retail and Manufacturing: the openings winners year over year (+155,000 and +152,000 respectively). Leisure and Hospitality lost 187,000 openings.
In Refolk's index of professional profiles, there are roughly 335,000 US software engineers at the Entry and Senior tiers, concentrated at Google, Microsoft, LinkedIn, Figma, and Glean. That is a supply pool bunched at name-brand ceilings that are not backfilling in July's report. Compare that to roughly 84,500 US business analysts, consultants, marketing managers, and financial analysts at the Senior and Entry-Manager tiers in the same index. Software engineering supply outweighs core white-collar management supply by about 4.0x. When employers post generic SWE reqs, they know the queue is deep, and they use that knowledge accordingly.
Postings became marketing. Your resume is answering an ad, not applying for a job.
The resume strategy for a slow hiring market
Stop optimizing for volume, start optimizing for authenticity and routing. In a market with 1.44 openings per hire and a 1.9% quits rate, the two levers that still move are picking real postings and getting your resume in front of the person who can actually hire, ideally before the req is even public.
1. Diagnose posting authenticity before you spend the hour
Before you tailor a single bullet, screen the posting. A quick filter:
- Age of the listing. Anything over 45 days at a mid-size company is suspect; over 90 days, treat as ghost by default.
- Salary range. Missing range in a jurisdiction that does not require one is a signal, not proof.
- Specificity of the JD. Generic "Software Engineer" or "Marketing Manager" with a boilerplate description at a 2,000+ person employer scores badly.
- Repost pattern. A req that closes and reopens monthly is a pipeline, not a hire.
- Recruiter activity on LinkedIn. If nobody on the team has posted about the role, nobody is hiring for it.
2. Route around the ATS when you can
Quits at 1.9% means the internal candidate pool at each employer is thin, and employers are backfilling through referrals. Targeted, warm-intro applications convert 5 to 10x better than portal submissions in a low-quit market. That does not mean skip Workday, Taleo, and Greenhouse entirely. It means the portal is the paperwork step, not the sourcing step. Find the hiring manager first, get referred, then submit.
3. Tailor at the posting level, not the role level
A generic resume is scored against a generic keyword set and loses to twenty other generic resumes. The only thing that moves you out of that pile is a resume that mirrors the specific verbs, tools, and outcomes in the specific JD. That is the exact work Refolk takes off you: paste the posting, get your own resume back rewritten for it, with a cover letter drafted and a fit score that tells you whether it is worth the send in the first place.
4. Cut the applications you were going to fake through
If your fit score comes back low, the win is not to force the application, it is to reallocate the hour. In a market with 4 hires for every 10 postings, low-fit applications are worse than nothing because they train you into the belief that effort equals output. Refolk's fit score exists to kill the applications you should not send, not to green-light the ones you should.
The cleanest single-month read on white-collar demand collapsing while postings hold.
Where the openings that are real actually sit
The growth is in physical and frontline work, not white-collar. Retail Trade added 155,000 openings year over year, Manufacturing added 152,000, and Leisure and Hospitality shed 187,000. If your resume assumes knowledge workers should chase "hot sectors," you have to grapple with the fact that the hot sectors are not knowledge work.
For a white-collar job seeker the practical read is:
- Accept longer search cycles. The market you are searching in has a 3.2% monthly hires rate.
- Skew toward AI-exposed hybrid roles. Indeed Hiring Lab's Nick Bunker has documented that AI-exposed occupations led the decline in postings from 2022 to 2026 and led the rebound over the past year. That is where the marginal white-collar opening is now sitting.
- Consider adjacent operations, manufacturing tech, or retail-tech roles that use your analytical skills against a growing sector rather than a shrinking one.
The referral math and the 23,400 "open to work" floor
Referrals convert because they route around the ghost-posting problem, not because they are magic. In Refolk's index, there are roughly 23,400 US professionals carrying an "open to work" signal in their headline, concentrated in the NYC and LA metros. That is a floor estimate of the actively searching white-collar pool. The math to internalize: every one of those people is applying to the same posting you are, and none of them get an interview unless someone inside the company argues for them.
Two moves follow from that:
- Build the referral graph before you need it. Every warm contact at a target employer is worth ten cold applications. Map who you know at each of your top 15 employers, not each of your top 150 postings.
- Time the outreach to the req. A hiring manager who just got budget approval will read your message. One who is three months into a ghost posting will not. Watch for headcount announcements, funding rounds, and internal promotions on LinkedIn.
The New York bill and what changes if it passes
A concrete policy hook is worth watching: New York State passed a bill requiring employers to disclose when they expect an advertised role to be filled and levying fines for violations. The bill was approved by both houses of the state legislature and, as of August 2026, was awaiting Governor Kathy Hochul's review. If it becomes law, the ghost-posting economics shift in New York specifically, and any candidate targeting NYC employers gets a cleaner signal on which reqs are real. Until then, the diagnostic work is on you.
FAQ
Is the 3.2% hires rate really comparable to 2010?
Yes, the arithmetic matches: the 3.2% July 2026 hires rate equals the 2010 annual average, per the JOLTS data BLS released on September 1, 2026. What is different is the unemployment context. In 2010, unemployment sat at 9.6% and the low hires rate was paired with mass layoffs. In 2026, unemployment is far lower because employers are not firing either, so the low-hire number describes mobility, not distress. For a job seeker, the mobility number is the one that determines how hard it will be to land your next role.
What is the fastest way to spot a ghost job in 2026?
Check age, salary, specificity, and repost pattern in that order. A posting over 90 days old at a mid-size employer, with no salary range and a boilerplate JD, that has closed and reopened once already, is a ghost with high confidence. Cross-reference LinkedIn for anyone on the team posting about the role. If the recruiter and hiring manager are silent, the req is a pool-builder, not a hire.
Should I keep applying through Workday, Taleo, and Greenhouse portals?
Yes, but treat the portal as the paperwork step, not the sourcing step. Portal-only applications convert poorly in a 1.9% quits market because employers are backfilling through referrals. Use the portal after you have already found and messaged a warm contact or the hiring manager. That inverts the usual funnel: identify the person first, submit the paperwork second, so your resume lands in the reviewer's inbox with a name attached.
How many applications should I be sending per week?
Fewer than you probably are, and each one tailored. In a market with 1.44 openings per hire and 1 in 7 postings ghosted, sending 50 generic applications a week produces less signal than sending 8 to 12 sharp ones. The right measure is not applications sent, it is fit-scored applications sent to postings you have diagnosed as real, ideally with a warm intro attached. That is the shape of a job search that converts in a frozen market.