The one dataset that exists, and what it counted
There is exactly one regularly published national source that tells you what sellers give buyers: the National Association of REALTORS Profile of Home Buyers and Sellers. The 2025 edition covers transactions completed between July 2024 and June 2025, and its finding on incentives is short enough to quote whole.
Only 27% of sellers offered any incentive at all to attract a buyer. Among all sellers surveyed:
- 11% gave assistance with closing costs
- 9% provided a home warranty policy
- 7% gave a credit toward remodeling or repairs
- 5% gave some other incentive
- 1% paid for a mortgage rate buydown
Read as a whole, the picture is that roughly three sellers in four gave nothing, and the single most common thing the remaining quarter gave was help with closing costs rather than money for repairs. That is the entire national evidence base. Everything else you will find on this subject is somebody's impression of their own market.
A sourcing note, because it affects how much weight the figures carry. The copy of the report consulted for this page is a state association's hosted mirror; NAR sells the report itself. The sample size is not stated in the excerpt obtained, which means the margin of error around each of those percentages is unknown.
What the 7 percent is not
This is the most misused number in the subject and the distinction is worth being pedantic about.
Seven percent is the share of all sellers who gave a repair credit. It is not the share of repair credit requests that succeeded. Those are different quantities with different denominators, and they are not close to each other.
Consider why. The 27% who offered incentives includes sellers who offered them unprompted, in listings, before any buyer existed - a rate buydown advertised to attract traffic is in that 27%. It also includes sellers whose buyers never asked for anything. And the 73% who gave nothing includes a large number of sellers who were never asked, because their buyer waived the inspection contingency, or found nothing worth raising, or simply did not try. Dividing the 7% by anything to produce a success rate requires knowing how many buyers asked, and nobody has counted that.
So when you read that 7% of sellers give repair credits, the correct inference is about how common the instrument is across the whole market, not about your odds. If you want an honest statement of your odds: no published source measures how often a post-inspection credit request is granted. Not a federal series, not a trade survey, not an academic paper. Anyone quoting you a success rate has made it up, and the research behind this page records that absence as a deliberate gap rather than something to paper over.
The asymmetry that is actually useful
There is one genuinely actionable thing in the NAR incentive data, and it is the ordering rather than any individual figure. Closing-cost assistance (11%) is given more often than a repair credit (7%), and a home warranty policy (9%) is given more often as well.
Treat that as information about what sellers are comfortable agreeing to. Closing-cost assistance is the familiar transaction; their agent has papered it a hundred times, it fits on one line of the amendment, it does not require anybody to admit the house has a problem, and it does not set a number that becomes a story about the property. A credit "toward repairs" is the same money with a label on it that invites an argument about whether the repair is really necessary and really costs that much.
The implication is not that you should hide what you want the money for. It is that the label on the request is doing more work than buyers realize, and that the instrument sellers say yes to most often is the one that reads as ordinary. If your constraint is cash at closing - which it usually is, since a credit and closing-cost assistance both land on the same side of the Closing Disclosure - the framing is free to change and the dollars are not.
A caution on the warranty line. A home warranty appears in the data more often than a repair credit, and it is cheap for a seller to grant, which is exactly why it gets offered in place of money. Whether that trade is good for you depends on what the policy actually covers, which is a different subject and belongs to a different site.
How much room the market leaves you
Two further figures set the scale, and both of them should lower your expectations.
From the same 2025 Profile of Home Buyers and Sellers: recent buyers "typically purchased their home for a median of 99 percent of the asking price." One percent is the median gap between what a seller asked and what a buyer paid. That median blends the original negotiation with whatever happened after an inspection, so it cannot be attributed to inspection findings and it is not a cap on anything. What it tells you is that the distribution of outcomes is tight, and a request in the high single digits as a percentage of price sits far outside it.
From NAR's August 2026 REALTORS Confidence Index survey - a monthly series, which is why the month has to travel with the number - 16% of homes still sold above list price, down from 19% the prior month and 20% a year earlier. Listings drew an average of 2.1 offers, down from 2.5 a year earlier. Median time on market was 31 days and contracts typically closed in 30.
Put those together and you have the seller's alternative, quantified. In a month when roughly one house in six goes for more than the asking price and the typical listing has had more than one offer, "we will put it back on the market" is a sentence with evidence behind it. Softening offer counts cut the other way and are the reason the same request lands differently this year than last. The direction of travel in those numbers is more useful to you than any single reading of them.
What the data says about buyers, which changes the seller's read on you
The same August 2026 Confidence Index reports that 20% of buyers waived the inspection contingency, up from 16% the prior month and 18% a year earlier, and that 22% waived the appraisal contingency. Cash sales were 27% of the market. First-time buyers were 30%.
Those numbers describe the field you are negotiating in. One buyer in five is going into these transactions with no inspection-based right to renegotiate at all, which means a seller who has recently sold or recently listed has a decent chance of having dealt with a buyer who asked for nothing. More than a quarter of sales involve no lender, and therefore none of the concession caps and none of the lender-required repairs that give a financed buyer leverage they did not have to create themselves.
The survey methodology, stated so you can weigh it: the August 2026 survey went to a random sample of 75,000 REALTORS plus 3,924 prior respondents who had given their email addresses; 1,718 responded, the survey ran from August 31 to September 3, 2026, and 872 respondents had a client in the last month. Among those with a client, the maximum margin of error is 3% at the 95% confidence level. These are agents reporting on their own recent transactions, not a count drawn from records.
What to take from this, and what not to
Three conclusions the data supports:
- Giving nothing is the normal outcome. Three sellers in four gave no incentive at all in the 2025 survey year. If you plan on the assumption that a request will be met, you are planning against the base rate.
- The form of the ask matters more than the data suggests the amount does. Sellers agreed to closing-cost help more often than to repair money. Same dollars, different line, measurably different frequency.
- The whole negotiation happens inside a narrow band. A median sale at 99% of asking is not a market where double-digit concessions are routine.
And the conclusions the data does not support, which are the ones circulating: that most sellers negotiate after an inspection, that a given percentage of requests succeed, that there is a typical credit amount, or that inspection findings are the usual reason a price moves. None of those are measured anywhere. The useful thing about knowing that is it stops you from calibrating against a number somebody invented.
For what determines whether a credit can be paid at all once a seller has agreed, see lender limits on seller credits. For which instrument to ask for in the first place, see seller credit versus price reduction.