“FSBO homes sell for 26% less than agent-listed homes”
The 26% figure appears in no current data source: ClozeEZ traced it to the machine-readable FAQ markup on one agent-referral company's page, contradicted by that page's own visible text. The National Association of Realtors' 2025 survey, the data this claim is loosely derived from, shows a raw median gap of 15.3%: $360,000 for FSBO sales against $425,000 for agent-assisted sales. That raw gap shrinks to 12.4% once you exclude FSBO sellers who sold to someone they already knew, which was 60% of them, and the only peer-reviewed study to compare like homes with like, published in the American Economic Review, found FSBO sellers obtained prices no lower than MLS sellers. The kernel of truth: raw FSBO medians really are lower. The myth is treating that raw gap as the price of skipping an agent.
Where does the 26% number actually come from?
Not from NAR, and this part is worth slowing down for. The figure is popularized by the statistics page of Clever Real Estate, a company that earns referral fees for matching sellers with agents. That page's own body text currently describes the gap as 'nearly 15%.' ClozeEZ found the '26% less' sentence surviving only in the page's machine-readable FAQ markup: the structured data that search engines and AI assistants scrape when they answer questions. In other words, the number that keeps resurfacing in AI answers and Reddit threads is a leftover in code, contradicted by the visible page it lives on.
No current edition of NAR's Profile of Home Buyers and Sellers supports 26%. The 2025 edition, published November 2025 and covering sales from July 2024 through June 2025, puts the raw medians at $360,000 and $425,000, a 15.3% gap.
What does NAR's own data actually show?
Three things the headline never carries. First, 60% of FSBO sellers sold to someone they already knew, a relative, friend, or neighbor, and NAR's own exhibit shows FSBO sellers who did not know their buyer had a median of $372,200, cutting the open-market gap to 12.4%. A pre-arranged sale to family at a family price is a different transaction from an open-market listing, and it is more than half the FSBO sample.
Second, the homes are different. NAR's own commentary notes FSBO homes are more often lower-cost manufactured homes or rural properties: 16% of manufactured-home sales were FSBO against 5% of sales overall, and the FSBO share runs 10% in rural areas versus 3% in suburbs. Comparing medians across different housing stock measures the stock, not the selling method.
Third, the detail that should change how you read all of it: 40% of FSBO sellers did not actively market the home at all, and 90% offered no incentives. The survey is not comparing marketed FSBO sales to marketed agent sales; a large share of its FSBO sample never tried. NAR's methodology is a mail survey of buyers (6,103 usable responses, a 3.5% response rate), not a controlled comparison of the same home sold two ways.
What happens when researchers compare like with like?
The gap mostly disappears. The one peer-reviewed study built for exactly this question, Hendel, Nevo, and Ortalo-Magné, published in the American Economic Review in 2009, compared homes sold through a strong FSBO platform in Madison, Wisconsin against MLS sales, controlling for property characteristics. FSBO sellers obtained sale prices no lower than MLS sellers, which means higher proceeds net of commission. The honest trade the study did find: FSBO took longer and carried a lower probability of selling, and patient sellers self-selected into it.
Two studies point the other way, and you should know what each actually measured. Collateral Analytics, an industry data firm, found FSBOs sold roughly 5.5% to 6% lower controlling for characteristics in 2016-17 data; not peer-reviewed, but a real counterpoint. And the widely cited Bright MLS/Drexel study claiming a 17.5% premium did not compare FSBO with agent sales at all: it compared on-MLS with off-MLS transactions in the Mid-Atlantic, where off-MLS includes agent-represented pocket listings and family transfers pulled from deed records, the same confounder as NAR's medians, and it was commissioned and published by an MLS.
So what is the honest takeaway for an Arizona seller?
The data does not say selling yourself costs you 26%, or 15%, or any fixed number. It says outcomes track behavior. The FSBO sellers dragging the median down include the 60% doing pre-arranged deals with people they knew and the 40% who never marketed. The sellers the Madison study describes, who priced from real comparable sales, photographed the home properly, and put it where buyers look, got full price and kept the commission.
FSBO share sits at an all-time-low 5% of sales, which makes the myth self-reinforcing: the smaller and stranger the sample, the worse its raw median looks, and the more the raw median gets quoted. If you sell yourself, the number that applies to you is not the median of people who never tried. It is the price your preparation earns.
Sources
- NAR 2025 Profile of Home Buyers and Sellers (full report): FSBO median $360,000 vs agent-assisted $425,000; 60% of FSBO sellers knew the buyer; unknown-buyer FSBO median $372,200; 40% did not actively market. Published November 2025.
- NAR: FSBOs reach all-time low: NAR's own framing, including its caveat that FSBO homes skew toward lower-cost manufactured and rural properties.
- Hendel, Nevo & Ortalo-Magné, American Economic Review 99(5), 2009: Peer-reviewed like-for-like comparison: FSBO-platform sellers obtained prices no lower than MLS sellers; FSBO took longer to sell.
- Bright MLS / Drexel 'On-MLS Study' (2023): The 17.5% figure: measures on-MLS vs off-MLS in the Mid-Atlantic, not FSBO vs agent; commissioned by an MLS.
- Clever Real Estate FSBO statistics page: The source of the viral 26%: the page's body text says 'nearly 15%'; the 26% sentence survives only in its FAQ schema markup.
Related questions
Is it true that most FSBO sellers already know their buyer?
Yes: 60% in NAR's 2025 survey, up sharply from 38% the year before. For 30% of FSBO sellers, selling to a relative, friend, or neighbor was the main reason they went FSBO at all. That is why raw FSBO medians say little about what an open-market FSBO sale fetches.
Do FSBO homes take longer to sell?
The best evidence says yes. The American Economic Review study found FSBO carried longer time-to-sale and a lower probability of selling, even though prices were no lower. Patience is the real cost the data supports, not a fixed price discount.
What did the Bright MLS 17.5% study actually measure?
On-MLS versus off-MLS transactions in the Mid-Atlantic, not FSBO versus agent. Off-MLS sales pulled from deed records include agent-represented pocket listings and non-arm's-length family transfers, and the study was commissioned and published by an MLS. It is a real study answering a different question.
So does selling FSBO cost you money or not?
It depends entirely on how you do it. The raw medians are dragged down by pre-arranged sales and unmarketed listings. The controlled evidence says a FSBO seller who prices from comparable sales and actually markets the home gives up little or nothing on price, keeps the commission, and pays for it in patience.