Ghostwriting Statistics (2026): Rates, AI and Disclosure

You are building a pitch, a budget or a board slide, and you need one number about ghostwriting. Maybe it is what a book ghostwriter costs. Maybe it is how many executives have someone else writing their LinkedIn posts. You search, and the first page hands you a dozen confident figures. “60 percent of nonfiction is ghostwritten.” “Ghostwriting is a multibillion-dollar industry.” Very few of them name a survey, a sample size or a year.

Use one of those in front of a skeptical CFO or a journalist and you are the one who looks careless. So I went looking for ghostwriting statistics I could trace to a named source, a method and a date, and threw out the rest. What survived is thinner than the listicles suggest, and more useful.

Why ghostwriting statistics are so hard to trust

Ghostwriting is confidential by design. Contracts often include nondisclosure clauses, the writer’s name rarely appears on the cover, and the client has every reason not to volunteer the arrangement. That means no regulator, publisher or trade body counts ghostwritten books or posts the way the Bureau of Labor Statistics counts editors. Gotham Ghostwriters says as much in its own 2026 pricing guide, noting that the industry has “no standardized pricing body” and has long “operated behind a veil of confidentiality.”

Into that vacuum step three kinds of publishers. Agencies and marketplaces publish data from their own transactions. Researchers publish studies on authorship, trust and AI that touch ghostwriting without measuring it directly. And content farms publish statistics pages that aggregate everything, including numbers that trace back to nothing.

The result is a field where the most-cited numbers often come from the companies that sell ghostwriting. That does not make them wrong. Gotham, Reedsy and the Editorial Freelancers Association all publish real data with stated samples. It does mean you should know who is talking before you repeat what they said.

Grade every number with the Ghost Grade

To sort what I found, I used a simple four-level scale I call the Ghost Grade. It rates a statistic on who produced it and how much of the method you can see, not on whether the number sounds plausible.

Grade A is peer-reviewed or methodologically transparent independent research, published with sample sizes and study design, by people with no commercial stake in the answer. Grade B is an industry survey with a disclosed sample and timeframe, run or sponsored by an organization that sells related services. Grade C is marketplace or vendor data drawn from a company’s own transactions or published as a range on a sales page, with limited method detail. Grade D is anything without a traceable source, sample or year. Grade D numbers do not appear in this post.

The grade travels with the number. When you quote a B, you name the sponsor. When you quote a C, you say it reflects one company’s marketplace. That one habit protects you from almost every embarrassment that ghostwriting statistics can cause, and I apply it to each section below.

What do book ghostwriters charge in 2026?

Rates are the best-documented corner of ghostwriting, because three organizations publish them with stated samples. They measure different things, which is why their numbers look so far apart.

The Gotham and American Society of Journalists and Authors survey, which Gotham called the first-ever survey of professional ghostwriters, sits at the premium end. Publishers Weekly reported its other findings in November 2024: one-third of respondents earn more than $100,000 a year from ghostwriting, half charged $10,000 to $20,000 for their last nonfiction book proposal, and 11 percent charged more than $20,000 for a proposal. Credentials drive the spread. Thirty percent of respondents had written at least one New York Times bestseller, and 2 percent had written ten or more, charging upward of $150,000 per manuscript, with 1 percent charging over $300,000. An agency and a professional writers’ association ran it, so treat it as a picture of the credentialed market, not of every freelancer on a gig platform.

Reedsy’s figure is a marketplace average, and it lands far lower. The EFA chart sits between them and covers more formats. Its 2026 medians for ghostwriting run 25 to 40 cents per word for blog posts, 15 to 30 cents for articles and essays, 50 cents to $1.00 for business and marketing writing, and 50 cents to $1.00 for memoir. The EFA stresses that these are reported medians for guidance, not recommended rates. Do the arithmetic on an 80,000-word nonfiction book at those medians and you get $26,400 to $40,000, close to Reedsy’s average.

Here is where hard data gets thin. Speechwriting has almost no independent rate data. The EFA chart folds speechwriting into an “Other Writing” catch-all rather than giving it its own row. The most specific public ranges come from Gotham’s 2026 pricing guide (grade C, since it is a seller’s sales page): $1,500 to $5,000 for a short or ceremonial speech and a conference keynote quoted at $10,000 to $45,000 in its text and $10,000 to $55,000 or more in its summary table. The same guide lists executive thought leadership retainers at $2,500 to $10,000 a month in its text and up to $20,000 in its table. When one source gives two ranges for the same service, quote the narrower one and say where it came from. Reedsy reports short-form collaborations (blog posts, op-eds, essays) at $500 to $2,000 in 2025, also grade C.

Does a ghostwritten book pay for itself?

This is the statistic people most want, and the one most often misquoted. The source is the 2024 Business Book ROI Study, conducted by author and former Forrester analyst Josh Bernoff with Amplify Publishing Group, Gotham Ghostwriters, Smith Publicity and the consultancy TLL. It surveyed 350 authors and prospective authors between April and August 2024, 301 of whom had published a nonfiction book. The report calls itself a non-probability survey, says sponsors did not review findings before publication, and recruited respondents through the sponsors and their partners. Grade B.

Its headline findings are useful. Across books out at least six months, median gross profit was $11,350, and 64 percent of business books showed a gross profit. The report says the median ghostwritten book was four times as profitable as other books. Among authors who paid for a ghostwriter, median spending on that service was $25,000, and 96 percent said they were satisfied, the highest satisfaction of any service in the study.

Now the misquote. Gotham’s FAQ and pricing pages describe $92,500 as the median gross profit for ghostwritten books. The underlying report’s charts show $92,500 as the median revenue for ghostwritten books and $43,250 as their median gross profit, which is what the report’s text calls “more than $43,500.” Revenue is not profit. If you cite this study, use the report, not a summary of it.

Also note the sample. Only about one author in twelve used a ghostwriter (5 percent alone, 3 percent with a coauthor), and only 25 authors reported ghostwriter costs. The report itself warns that “it is unwise to draw conclusions from statistics for categories with fewer than 25 respondents.” The ghostwriter group sits right at that line. And the study shows correlation, not cause. Authors who can afford a ghostwriter may also be better positioned to monetize a book through speaking and consulting, which the report found produced more income than book sales.

How common is ghostwriting on executive LinkedIn?

I could not find a credible, current, published survey measuring what share of executives use a human ghostwriter for LinkedIn. Figures like “80 percent of Fortune 500 CEOs” circulate on agency blogs without a named study behind them. They are grade D, and I left them out. If you find a figure with a sample size and a fielding date, you will be ahead of nearly everyone writing about this.

What exists is adjacent data. The Edelman and LinkedIn B2B Thought Leadership Impact Report surveyed 3,484 business executives across seven countries for its 2024 edition, fielded November 30 to December 14, 2023. Thirty percent said the overall quality of the thought leadership they read is mediocre, and only 13 percent rated it very good or excellent. That is not a ghostwriting statistic, but it explains the market. Executives are publishing at volume, readers find most of it weak, and that gap is what ghostwriters, editors and AI tools are being hired to close. Edelman and LinkedIn are grade B here: a disclosed method, run by a platform and an agency with an interest in thought leadership mattering.

The other adjacent data is AI detection, which measures something different from ghostwriting. A post written by a human ghostwriter will not be flagged by a detector, and an executive who drafts with ChatGPT has no ghostwriter at all. Keep those categories separate when you write about them.

AI-assisted ghostwriting by the numbers

The best data on how professional writers use AI comes from Gotham Ghostwriters’ 2025 AI and the Writing Profession study, designed with Josh Bernoff. MarketingProfs, summarizing it in November 2025, reported a sample of 1,190 writing professionals, while Gotham’s 2026 pricing guide describes 1,481. I have not been able to reconcile the two, so if you cite the study, give the source of the figure you use. Grade B.

Per MarketingProfs, 26 percent of professional writers use AI tools daily for writing, editing or research, 19 percent at least weekly and 16 percent sometimes, while 39 percent were classified as nonusers. Thought leadership writers were the most likely to use AI daily, at 33 percent. Gotham’s pricing guide adds that book ghostwriters use AI at a rate of 68 percent, and that only 7 percent of professional writers publish AI-generated text without significant human editing.

On the output side, the detection companies supply the numbers, and both have a commercial interest in finding AI. Originality.ai analyzed 8,795 long-form LinkedIn posts from January 2018 to October 2024 and estimated that 54 percent of long-form posts were likely AI-generated by October 2024. Its monthly tracker reported 75.6 percent for July 2026, based on 93 of 123 randomly selected posts, and an essentially unchanged share for August from 119 posts. Pangram, a rival detector, reported in July 2026 that more than 40 percent of long-form LinkedIn posts it examined were flagged as fully AI-generated, and that two-thirds of all posts it flagged came from LinkedIn.

I would grade these C. The methods are described, but detectors make errors, the monthly samples are small, and “likely AI” is a probability label, not a confession. They support a narrow claim (a large share of long-form LinkedIn text reads as machine-written) and not a broad one (most executives use AI ghostwriters). LinkedIn itself said in 2026 that it is reducing distribution of generic, repetitive content that appears AI-generated, according to Originality.ai’s report on the change, which gives the statistic practical weight for anyone ghostwriting on the platform.

Do readers punish disclosure?

This is where the grade A research lives, and it is not reassuring. In “The transparency dilemma: How AI disclosure erodes trust,” published in Organizational Behavior and Human Decision Processes in 2025, Oliver Schilke and Martin Reimann ran 13 experiments with more than 5,000 participants. In every one, people who disclosed AI use were trusted less. The University of Arizona’s summary gives examples: trust dropped 16 percent when students learned a professor used AI for grading, 18 percent when investors saw AI use disclosed in ads, and 20 percent for graphic designers. Softer wording, such as saying AI was used only for proofreading, did not remove the penalty. And the paper found the drop was larger when AI use was exposed by a third party rather than self-disclosed.

A 2026 study by Zhu and Molnar in Computers in Human Behavior, described by Andras Molnar in The Conversation, recruited more than 1,300 U.S. participants and showed them AI-generated personal messages such as an apology. When authorship was disclosed as AI, recipients rated senders as “lazy” and “insincere.” When nothing was disclosed, their impressions were as positive as when they were told a human wrote it. Most people, in other words, do not suspect AI unless they are told.

The finding most relevant to human ghostwriting comes from a 2026 paper in the Journal of Consumer Behaviour led by Danielle Hass. Across five preregistered experiments, people felt guilt when presenting AI-written heartfelt messages as their own, driven by perceived dishonesty. The researchers found the same guilt when a message was secretly written by a friend, but not when it came from a pre-printed greeting card. The discomfort is not about machines. It is about passing off someone else’s words as yours in a context where readers assume you wrote them.

On the writer side, the Authors Guild’s May 2023 survey found 91 percent of authors believe readers should know when AI has created all or portions of a work. That is a writer’s view, not a reader’s, but it shows where the profession’s norms point.

Put together, the grade A evidence says disclosure costs trust, concealment carries a larger cost when it is discovered, and context matters. Business readers have long accepted that executives get writing help. Personal, emotional content is where the penalty bites hardest.

Quote these numbers without getting caught out

Most ghostwriting statistics fail in the retelling, not the research. Run every figure through the Ghost Grade before it goes in a deck. Name the source and year in the same sentence as the number. For a grade B or C figure, name the sponsor or the marketplace. Prefer the original report to anyone’s summary of it, including the sponsor’s, because the ROI study shows how quickly revenue becomes “profit” in retelling. And when a figure has no traceable sample, say the data does not exist rather than reaching for the most quotable guess.

The short version: book ghostwriting rates are well documented, running from about $36,000 for an 80,000-word book on Reedsy to six figures for credentialed writers, with EFA medians in between. Ghostwritten business books correlate with higher profit in one small, sponsor-recruited sample, at a median of $43,250, not $92,500. No credible survey measures how many executives use human ghostwriters on LinkedIn. AI use among professional writers is widespread and mostly edited, detector estimates of AI text on LinkedIn run from 40 to 76 percent depending on who measures, and the best experimental research says disclosure costs trust while exposure costs more.