You have maybe 100 good tailored applications in you before burnout sets in. The Columbia Law Review's November 2025 JOLTS analysis says roughly 60 of them will vanish into government listings that never hire, and 48 of every 100 IT postings will do the same. This is a routing problem, not a motivation problem.
What the Columbia Law Review actually found
Ghost-job rates published November 13, 2025 by the Columbia Law Review Forum peg government at ~60%, education and health services at ~50%, and IT at ~48%. The paper defines a ghost job as a listing by a real company for a position that does not exist or that the employer has no present intention to fill.
The mechanism is boring and structural. Hires-per-posting has halved since 2019, from roughly 8 hires per 10 postings to 4. BLS JOLTS recorded 7.3M openings against 5.1M hires in July 2026, a gap that has averaged about 2.2M per month since early 2024. Senator Ruben Gallego wrote BLS Deputy Commissioner William Wiatrowski in June 2026 asking whether ghost postings are distorting the labor-market signal the Federal Reserve uses to set rates. When a sitting senator is worried about your applications skewing monetary policy, the problem is no longer anecdotal.
ResumeUp.AI's September 2025 crawl of U.S. LinkedIn found 27.4% of listings show ghost-job characteristics. Clarify Capital's January 2025 survey of 1,000 employers found 81% of recruiters admit their company has posted a ghost job. ResumeBuilder's 2024 survey of 1,600 hiring managers puts the company-level rate at 40% in the past year.
Columbia Law Review Forum, November 2025, analyzing BLS JOLTS data.
The sector map, and the "real hire rate" you should be routing to
Sort every posting you consider by the inverse of its published ghost rate, apply hardest where that number is highest, and treat the sector rate as a prior you update with posting-level signals.
| Sector or slice | Published ghost rate | Refolk's index, U.S. incumbents in role | Implied real-hire rate |
|---|---|---|---|
| Government (all) | ~60% (Columbia Law Review) | 15,300 Management, Program, and Policy Analysts in gov and public-policy employers | ~40% |
| Education and Health Services | ~50% (Columbia Law Review) | not queried | ~50% |
| Information Technology | ~48% (Columbia); 47% vacancy gap with 86% candidate-ghosted (Enhancv 2026) | 330,856 U.S. "Software Engineer" incumbents | ~52% |
| LinkedIn, all U.S. listings | 27.4% (ResumeUp.AI, Sept 2025) | - | ~72.6% |
| Senior-level postings | ~20%, roughly 1 in 5 (Clarify Capital 2026) | - | ~80% |
| Employer-side self-report | 81% of recruiters admit posting ghosts (Clarify Capital, Jan 2025) | - | - |
The counter-intuitive line in that table is the last useful one: senior postings are roughly 2.4x less likely to be ghosts than IT postings overall. If you are mid-career and reflexively applying to more junior titles to widen your funnel, you are pushing toward the noisier end of the market, not the safer one.
Why government's 60% is a compliance artifact, not a scam
Government ghost postings are largely legal requirements to advertise externally when an internal candidate is already lined up. The 60% is high but predictable, and the filter is public and knowable.
Federal and state hiring rules force many roles to be publicly listed even when an internal transfer or a specific preferred candidate has been identified. For USAJobs postings, the "area of consideration" field and the "who may apply" section tell you whether the listing is open to the public or restricted to current federal employees, veterans, or a specific agency. In Refolk's index, about 15,300 U.S. incumbents currently hold titles like Management Analyst, Program Analyst, or Policy Analyst inside employers such as the U.S. Department of Veterans Affairs, FEMA, FDA, the City of San Antonio, and the Texas Alcoholic Beverage Commission. Those are the pools ghost-listed roles are usually being filled from.
Concrete filters before you tailor a federal or state resume:
- Check "area of consideration" on USAJobs. If it says "Status Candidates" or lists a single agency, external applicants rarely convert.
- Look at closing date windows. Postings open for 5 days or less are almost always pre-wired.
- Search the announcement number on LinkedIn. If the same number was posted 6 months ago and refilled the same title, it is a rolling requisition, not a hire.
- For state and municipal roles, check whether the posting names a working title that matches a person already at the agency on LinkedIn. That is your preselected internal.
Tailoring a federal resume takes 90 minutes at minimum because the format demands GS-level accomplishment statements with hours-per-week and supervisor contacts. Spending that on a Status-Only announcement is the single most expensive mistake in the government funnel. This is where Refolk earns its keep on the mechanical side: paste the posting, get your own resume back rewritten to the announcement's language, and get a fit score before you commit the 90 minutes of self-editing.
Why IT's 48% is the opposite problem: active pipelining
Tech's ghost rate is driven by talent-pool building, not compliance. In a labor pool of 330,856 U.S. Software Engineers in Refolk's index, the marginal resume costs employers nothing to collect, and Enhancv's 2026 read pairs a 47% IT vacancy gap with an 86% candidate-ghosting rate and calls it "performative hiring rather than simple delay."
The mechanism is that top employers in that pool - Google, Microsoft, Figma, Glean, and Ashby among them - keep evergreen requisitions open for pipeline. The requisition is real. The intent to hire this quarter is not.
Signals a tech posting is a pipeline listing, not an open seat:
- Posted more than 30 days ago with no "reposted" tag change.
- The title is generic ("Software Engineer") with no team, product, or stack named in the top third of the JD.
- The company has 40+ open engineering reqs but shipped no headcount announcement in the last two earnings calls.
- The recruiter listed has "Talent Community" or "University Recruiting" in their title.
- The salary band spans more than $80K, a sign the req covers 2-3 different levels the team has not decided between.
Mass-applying is negative-sum for all applicants, not just wasteful for you. Real applications crowd the queue for real openings.
The 100-application budget, allocated
Given the published sector rates, an IT candidate sending 100 tailored applications to generic LinkedIn IT postings should expect only about 52 to sit on top of a real hiring intent, and only a fraction of those will surface their resume. Reallocating that budget by seniority, source, and posting age is worth more than any resume rewrite.
A defensible split for a mid-to-senior IT candidate:
- 40 applications to senior-level postings less than 14 days old, at companies that announced headcount plans in their last earnings call. Ghost exposure: ~20%.
- 25 applications through warm referrals, where the ghost-rate question is moot because a human is routing you. Ghost exposure: effectively 0%.
- 20 applications to postings on company career sites (not LinkedIn aggregators), less than 21 days old. Ghost exposure: closer to the 27.4% LinkedIn baseline once you strip repost noise.
- 10 applications to public-sector or contractor roles where you can confirm "open to the public" status. Ghost exposure: ~40% instead of the government-wide 60%.
- 5 speculative applications to companies you would take a pay cut to join. Ghost rate irrelevant; you are buying a lottery ticket.
The tailoring math changes when you cut the funnel this way. Instead of 100 shallow customizations that get chewed up by 48% ghost noise, you write 75 pointed at postings with real hiring intent and 25 that route around the posting layer entirely.
20% ghost rate at senior level (Clarify Capital 2026) versus 48% for IT overall (Columbia Law Review).
Regulatory pressure that will move the numbers in 2026-2027
Ghost jobs are becoming a legal risk for employers, not just a candidate complaint. Ontario's Working for Workers Act hiring-disclosure provisions took effect January 1, 2026, and California and Kentucky passed similar statutes in 2025.
The Congressional Research Service brief IF12977 (2025) formally acknowledged ghost jobs while noting BLS and DOL do not officially track them, which is exactly why private estimates diverge so widely (27.4% on LinkedIn versus 60% in government). Expect a lag: legislation moves the honest employers first, and the ghost rate in government and IT will not drop meaningfully until federal disclosure rules follow. For 2026 and most of 2027, the sector map above is your operating reality.
How to spot ghost jobs before you tailor
A listing is likely a ghost if it fails at least two of these five checks, and you should not spend more than 10 minutes on it if it fails three.
- Age. Posted more than 30 days ago with no substantive edit. LinkedIn's "reposted" tag resets the clock but not the intent.
- Named humans. No hiring manager, no recruiter, no team lead visible on the posting or in a same-week LinkedIn post from the company. Government roles get a pass here.
- Requisition specificity. No team, product, stack, or reporting line named. Ghost postings read like job families, not jobs.
- Company motion. Recent layoffs, hiring freezes announced on earnings calls, or a headcount cap in the most recent 10-Q. Nearly 78% of Fortune 500 firms use an ATS, and Workday holds ~45% of that segment; ATS-first postings sit stale longest.
- The mirror test. Search the exact job title plus company on Google with a 6-month date filter. If the same requisition surfaces from January, April, and September, it is a pipeline listing.
Running this checklist on every posting is exhausting, which is precisely why most applicants skip it and burn their tailoring budget on noise. Feeding a posting URL into Refolk and getting back a fit score plus a rewritten resume, or a "skip this one" signal, is the version of this that scales past 20 applications a week without turning into a second job.
FAQ
Are ghost jobs illegal?
Not in most U.S. jurisdictions yet. Ontario's Working for Workers Act, effective January 1, 2026, and California and Kentucky's 2025 statutes are the first meaningful disclosure regimes, but federal enforcement does not exist and BLS does not officially track ghost postings per the Congressional Research Service's 2025 IF12977 brief. Employers face reputational risk, not legal risk, in most states.
If 27.4% of LinkedIn postings are ghosts, why is the IT rate 48%?
Different denominators. The 27.4% ResumeUp.AI figure covers all U.S. LinkedIn listings across every sector, weighted heavily toward sales, retail, and healthcare, where ghost rates are lower. The 48% Columbia Law Review number isolates the IT sector using BLS JOLTS data, where evergreen pipelining by large tech employers is standard practice. Both numbers can be true simultaneously.
Should I stop applying to government jobs?
No. The 60% ghost rate in government is largely a compliance artifact, meaning postings are legally required to be public even when an internal candidate is preselected. Screen USAJobs postings by the "area of consideration" field, avoid Status-Only announcements as an external candidate, and treat closing windows shorter than 5 days as pre-wired. The remaining 40% is real, and government roles convert at high rates once you clear the screening.
Does tailoring a resume even matter if half the postings are ghosts?
Yes, for the real half, more than ever. Ghost postings still receive real applications, which crowds the queue for actual openings, so a generic resume loses twice: once to the ghost tax and once to sharper competitors on the real reqs. The rational move is not less tailoring, it is less tailoring wasted, which means scoring postings for hire intent before you invest the 40 minutes.