Ghost Jobs in 2026: Why 1 in 5 Senior Engineering Postings Isn't Real

You have been applying for eleven weeks. Your résumé is good -- you have shipped things people have heard of, you can talk about consistency models without notes, and three former colleagues would take your call at midnight. And yet the funnel behaves strangely. Applications go in. Nothing comes back. Not rejections. Nothing. You check a posting a month later and it is still there, still "actively recruiting," still wearing the same three-sentence description it had in May.

The instinct is to blame yourself, then the résumé, then the market. There is a fourth explanation almost nobody prices in correctly, and it has the best data behind it: a substantial share of the jobs you applied to were never going to be filled by anyone.

This is not a conspiracy theory. It is a measurable, surveyed, now-legislated phenomenon that lands disproportionately on people at your level. Here is what the evidence shows, why senior and staff engineers are the most exposed cohort, and how to run a cheap verification pass that keeps phantom reqs from eating a quarter of your life.

The Statistical Fingerprint of a Market Full of Phantoms

Start with the macro picture, because it is genuinely strange. The Bureau of Labor Statistics' Job Openings and Labor Turnover Survey for May 2026 counted 7.6 million open positions. In the same month, employers made 5.2 million hires. Openings have drifted upward alongside payroll growth; hiring has not moved with them.

A gap between openings and hires is normal; some fraction of every posting is mid-process. What is not normal is the gap persisting month after month while the quits rate signals that workers have stopped believing they can move. Some of that divergence is friction, some is employer caution, and some -- nobody knows how much -- is postings never attached to a hiring decision at all. The Congressional Research Service now maintains a briefing on "ghost" job postings partly because they degrade the reliability of the openings data itself.

Your specific market has its own shape. Software development postings remain roughly 27.5% below their pre-pandemic level, though they have grown about 15% since early 2025. The part that matters for you: 71% of the increase between May 2025 and May 2026 came from senior roles, and 37% from jobs with AI in the title. The demand that exists is concentrated at your level -- which is also where the phantoms concentrate.

Five Species of Ghost Job (Only Two Are Malicious)

"Ghost job" gets used as a single accusation. It is really five distinct behaviors with different implications for how you respond.

1. The stale req. The role was filled or frozen and nobody took the posting down. Boring, extremely common, and not evidence of bad faith -- just of a recruiting org drowning in volume.
2. The evergreen pipeline. A standing posting for a role the company hires continuously ("Senior Backend Engineer -- Multiple Locations"). Sometimes real, sometimes a bucket. Applications land in a pool, not a process.
3. The internal-candidate formality. The role is spoken for, but policy requires a public posting. Your application is compliance paperwork.
4. The résumé harvest. No req, no budget, no timeline. The posting exists to build a candidate database for a future that may not arrive. This one is deliberate.
5. The signaling post. The posting exists to make the company look like it is growing -- to investors, to competitors, or to its own exhausted staff. Also deliberate, and the most cynical of the five.

Categories one through three cost you an application. Four and five can cost you three interview rounds, a take-home, and six weeks of hope -- which is what the verification protocol below is built to catch.

Why Senior Engineers Get Hit Hardest

The most rigorous recent measurement comes from a February 2026 analysis by Clarify Capital, which scraped 176,268 unique live listings on Indeed across 49 industries and every U.S. state, flagging any posting that had stayed active 30 days or longer. Their headline: roughly one in seven active postings qualified, and 4% had been live for four months or more.

Greenhouse, which runs the applicant tracking system behind a large share of tech postings, reports a similar band from the inside: 18–22% of jobs posted on its platform in any given quarter are ghost jobs, and 60% of candidates in its 2,500-person survey suspected they had encountered one. In its 2026 workforce report, drawn from 2,200 active job seekers, only 7% believed the market favored them and 72% said the job they were eventually offered differed from the one they applied for.

Why does seniority concentrate the problem? Senior reqs are the ones most often opened speculatively -- a director wants to see who is available before committing headcount. They are the ones most often frozen mid-process when a budget cycle turns, and a frozen req usually stays posted. And they carry the most leveling ambiguity: a company unsure whether it wants a Senior, a Staff, or nobody will leave the door open and see who walks through. That same fuzziness is why so many candidates get down-leveled at the offer stage -- phantom reqs and phantom levels come from one source.

The 30-Day Test Fails at Your Level

Here is the analytical trap almost every "how to spot a ghost job" article walks into, and it matters more for you than for anyone else.

Nearly all ghost-job research uses posting age as its proxy: 30 days live, therefore suspicious. But Ashby's May 2026 recruiting-operations benchmarks -- built on more than 54 million applications across 93,000 jobs -- put the median time to first fill for senior roles at 71 days, versus 63 for mid-level and 52 for junior. Senior roles take 37% longer to fill than junior ones.

This cuts both ways. The widely-quoted 21% figure is almost certainly an overcount of genuine bad faith, because normal senior hiring is slow. It also means you cannot use the one heuristic everyone recommends. A 45-day-old staff engineer posting is unremarkable. A 45-day-old posting where the recruiter cannot name the hiring manager is a phantom. The difference is not in the timestamp; it is in whether anyone on the other side is behaving like a person with a deadline.

What Hiring Managers Admit in Their Own Surveys

Candidate-side data can be dismissed as bitterness. Employer-side data cannot, because employers volunteered it about themselves.

ResumeBuilder.com surveyed 1,641 hiring managers in May 2024 and found 40% said their company had posted a fake job listing in the prior year and 30% had one live at the time of the survey. The stated reasons, in order: to appear open to external talent (67%), to suggest the company was growing (66%), to create the impression that overworked employees' load would soon ease (63%), to make existing employees feel replaceable (62%), and to collect résumés (59%).

The number that should change your behavior is further down. Among companies that contacted applicants for these postings, 85% conducted interviews anyway -- with no intention of hiring. That is not a wasted click. That is a recruiter screen, a hiring manager conversation, possibly a system design round, extracted from you for a role that does not exist.

An earlier survey of 1,045 hiring managers by Clarify Capital found the same pattern from the other direction: half kept postings up because the company is "always open to new people," a third to be ready "in case an irresistible candidate applies." Neither is exactly a lie. Both mean the same thing for your calendar.

Layer on the volume problem. LinkedIn now sees roughly 11,000 applications submitted per minute, up about 45% year over year, much of it AI-assisted, and 22% of Greenhouse's respondents admitted to using bots to apply automatically. Recruiters answered in kind: 66% plan to expand AI pre-screening in 2026, and 93% of talent acquisition professionals say they will grow their AI usage. A phantom posting can now absorb thousands of machine-generated applications at zero cost to the employer -- which is precisely why the incentive to take one down is so weak.

The Real Cost: Funnel Math, Not Feelings

Do the arithmetic on your own search rather than absorbing this as vague discouragement.

A properly tailored senior application -- reading the posting, adjusting the résumé, writing something that is not boilerplate -- costs 30 to 45 minutes. Send 60 over a quarter and that is roughly 35 hours. If 20% are phantoms, you have donated seven hours to reqs that do not exist. Painful, but survivable.

The interview leakage is where it gets expensive. Ashby's benchmarks put overall screen passthrough at 35% and onsite passthrough at 24%. If a phantom req pulls you into a recruiter screen plus a hiring manager round, that is roughly four hours of live time plus preparation. Two of those in a quarter and the phantom tax approaches 20 hours -- a real interview loop you never got to run.

The second-order cost is worse and less visible: you cannot calibrate. If a fifth of your applications went into a void, your response rate is not a measurement of your résumé's quality. Engineers respond to a low response rate by rewriting the résumé, then rewriting it again, then concluding the market has passed them by. That inference is invalid when the denominator is contaminated. Before you tear apart a résumé that may be working fine, clean the denominator.

Nine Signals That Separate Real Reqs From Résumé Harvests

No single signal is conclusive. Two or three together, at a level where you know normal hiring takes 71 days, usually are.

1. Repost cadence. A role that reappears with a fresh date every two or three weeks for months is not being filled. LinkedIn's "Reposted" label is the cheapest tell available.
2. No requisition ID. Real internal reqs carry identifiers, and they usually survive into the ATS URL. A posting with no traceable req is a posting nobody is accountable for.
3. Aggregator-only presence. If the role is on Indeed and ZipRecruiter but not on the company's own careers page, you may be looking at a scraped, stale, or invented listing.
4. A description that has not changed. Live reqs get edited -- scope narrows, a language gets added, the level shifts. A description that is byte-identical after nine weeks is a description nobody is managing.
5. Implausible band width. A $150K–$320K range for one role is either three roles or no role.
6. Level soup. "Senior/Staff/Principal Engineer" in a single posting means the company has not decided what it is buying -- which frequently means it has not decided whether to buy.
7. No adjacent hiring activity. One senior posting at a company with no other open roles, no recruiter outreach, and no engineering hires on LinkedIn in six months is an outlier worth distrusting.
8. The recruiter cannot name the hiring manager. This is the strongest single signal in the list. A real req has an owner with a headcount problem.
9. No timeline, ever. Real processes have dates. "We're moving quickly" repeated across three weeks without a scheduled next step is the sound of a pipeline with no destination.

The Four-Check Verification Protocol

Run this before writing a tailored application. It takes about ninety seconds, and it is the highest-leverage ninety seconds in a senior job search.

Check 1 — Go to the source. Find the role on the company's own careers page, which almost always runs on Greenhouse, Lever, or Ashby. If it is not there, stop; aggregator listings routinely outlive their reqs by months. This is why pulling roles directly from ATS endpoints beats browsing a job board -- you are reading the employer's live state, not a cached copy of it.

Check 2 — Establish the posting's history. Compare the ATS listing date against the aggregator date. A three-month-old ATS req reposted to LinkedIn every fortnight is a company gaming freshness, not a company hiring.

Check 3 — Look for a trust signal. Greenhouse issues Greenhouse Verified badges to companies with a demonstrated record of communicating with candidates -- a status that cannot be bought -- and surfaces those roles through MyGreenhouse. Not a guarantee the req is real, but a meaningful prior on whether you will be treated like a person.

Check 4 — Ask one question before you invest. On any first recruiter call: "Is this a backfill or new headcount, and has the req been approved in this quarter's budget?" Real recruiters answer instantly, because they know. A vague answer to a specific, ordinary, non-confrontational question is the most reliable phantom detector that exists. Fold it into your standard recruiter screen preparation and it costs you nothing.

You're Already in a Ghost Pipeline. Now What?

Sometimes you find out mid-process. The tells are consistent: scheduling gaps that stretch from days to weeks, an interviewer who has clearly not read your résumé, questions that probe your current employer's architecture more than your capabilities, and a hiring manager who talks about the role in the conditional tense. Three moves, in order.

Ask for the decision timeline in writing. "What does the rest of the process look like, and when do you expect to decide?" in an email creates a record and forces specificity. A real process produces a date. A phantom process produces adjectives.

Cap your investment. Decline take-homes over four hours for a req you have not verified. This is defensible at senior level and a useful filter in itself: employers running real processes usually negotiate scope, and the response tells you which kind you are in. The same logic applies to any unpaid take-home assignment.

Keep the pipeline moving anyway. Ashby's data shows the median interviewed candidate is archived 21 days after their last interview -- silence, not rejection, is the standard ending. With 61% of job seekers reporting they were ghosted after an interview, plan for no answer as the base case; our guide on what to do when you get ghosted covers the follow-up sequence.

Rebuilding the Funnel Around Reqs That Exist

The strategic response to a market with a 20% phantom rate is not to apply harder. It is to change the channel mix, because different channels have wildly different phantom exposure.

Referrals are the strongest correction available, and the data is unambiguous: Ashby found 52% of referred candidates pass the initial screen versus 35% overall, with referred technical candidates accepting offers at 84%. Higher passthrough matters, but the bigger benefit is structural -- a referral is proof the req exists, because a human inside the company confirmed it before you spent an hour.

Recruiter-sourced roles carry a similar guarantee for the same reason: nobody sources for a job that isn't open. Sourced technical candidates accept at 76%. Inbound applications to public postings are the only channel with meaningful phantom exposure, so the correct posture is not "stop applying" but "stop treating the public posting as your primary channel."

Two adjustments follow. Track your funnel by channel, not in aggregate: if referrals convert at 40% and cold applications at 2%, that is not a message about your résumé, it is a message about where your hours belong. And track posting provenance -- log whether each role came from a company ATS, an aggregator, a referral, or recruiter outbound, and after 30 applications you will have a personal ghost-rate estimate more accurate than any survey. It is the same bookkeeping that tells you whether jumping or staying is the better financial move.

The Law Is Finally Catching Up

For most of the last decade, posting a job you had no intention of filling carried no consequences. That changed in 2026.

Ontario's job posting rules took effect January 1, 2026. Employers with 25 or more employees must now state in every public posting whether it is for an existing vacancy, disclose whether AI is used to screen applicants, and include a compensation range. Penalties run to CAD $100,000.

In the United States, New York moved first. Senate Bill S8877, sponsored by Senator Michael Gianaris, passed the Senate 39–19 on April 30, 2026 and cleared the Assembly on June 2. It sorts every posting into one of three buckets: filling within 90 days with a stated date, filling later with "no sooner than" language, or résumé collection only. Violations carry $2,500 penalties that double every 30 days. A companion bill would force employers to notify job boards when a role is filled, with removal required within a week.

Forbes counts five states with active ghost-job bills -- Pennsylvania, New Jersey, New York, California, and Kentucky. Pennsylvania's HB2321 is the most aggressive, adding mandatory removal within two weeks of a fill and limits on how long employers may retain applicant data. Bloomberg Law reads this as the successor to the pay transparency wave: the same regulatory logic, applied to a different piece of withheld information, for the same reason -- employers would not disclose it voluntarily.

None of this helps you this quarter. It does tell you the problem is real enough that five legislatures independently decided it needed a statute -- worth remembering the next time you are tempted to read a silent funnel as a verdict on your abilities.