Existing-Home Sales Explained: What SAAR Means—and What the Monthly Report Actually Measures

Existing-home sales measure completed transactions on existing single-family homes, condominiums, and co-ops. The figure comes from the National Association of REALTORS® (NAR), which publishes it monthly in its Existing-Home Sales report.

The headline number is easy to misread. In its October 13, 2026 release, NAR reported that existing-home sales rose 1.5% from August to September 2026, to a seasonally adjusted annual rate of 3.99 million. Sales were 2.2% lower than in September 2025. (Source: NAR, Existing-Home Sales release, October 13, 2026.)

That 3.99 million does not mean 3.99 million homes sold in September. It is a seasonally adjusted annual rate, or SAAR. The figure takes September's sales pace, adjusts it for normal seasonal patterns, and expresses it as the total that would result over a full year if that relative pace continued for 12 months. It is a measure of pace, not a monthly count, and it is not a forecast of the coming year.

This article explains what counts as an existing-home sale, what seasonal adjustment and annualization each do, how NAR builds the estimate, why published figures can be revised, and how existing-home sales differ from pending sales, inventory, and prices. It contains no forecast.

About the data

This article uses a small number of NAR's headline figures, with attribution, to explain how the indicator works. It does not reproduce NAR's data tables. NAR restricts redistribution of its Existing-Home Sales data.

What counts as an existing-home sale

NAR describes existing homes, unlike new homes, as homes that were owned and occupied before coming onto the market. Its Existing-Home Sales data are a monthly measure of the sales volume and prices of existing single-family homes, condos, and co-ops, nationally and for four regions: the Northeast, Midwest, South, and West. (Source: NAR, Existing-Home Sales Explained.)

Three features define what the report counts.

• It counts completed transactions. NAR says the monthly data capture completed transactions, and its release notes state that existing-home sales are based on transaction closings from Multiple Listing Services. (Sources: NAR, Existing-Home Sales Explained; October 13, 2026 release.) A sale enters the data when it closes, not when a contract is signed.
• It covers existing homes. The report covers single-family homes, condominiums, and co-ops that have been previously owned and occupied. NAR's release notes also list townhomes among the property types included. (Source: NAR, October 13, 2026 release.) Sales of newly built homes are measured separately.
• It reflects MLS activity. Because the data come from MLS closings, NAR notes that changes in sales trends outside MLSs are not captured in the monthly series, though NAR periodically benchmarks home sales against other sources. (Same source.)

Contract signed→

Pending→

Closing→

Existing-home sale

Existing-Home Sales measure completed transactions; Pending Home Sales measure earlier contract activity.

NAR also notes that existing homes account for more than 90% of total home sales. (Same source.) The report is therefore a measure of most, but not all, home-sale activity, and specifically of the resale segment.

The key practical distinction is simple: existing-home sales are not new-home sales.

Why the headline says 3.99 million when 3.99 million homes did not sell in September

3.99M SAAR

Seasonally adjusted+Annualized pace

×Not 3.99M homes sold in September

×Not a forecast for the next 12 months

A common misreading of the report is to treat the headline as a monthly count. It is not.

The 3.99 million figure is a seasonally adjusted annual rate (SAAR). NAR explains that the annual rate for a particular month represents what the total number of sales for a year would be if that month's relative pace were maintained for 12 consecutive months. (Sources: NAR, Existing-Home Sales methodology; October 13, 2026 release.)

Two separate steps are packed into that phrase:

• Seasonally adjusted: the month's sales are adjusted to remove the normal seasonal swing in home sales, so that September can be compared fairly with August or any other month.
• Annual rate: the adjusted monthly pace is expressed as the total that would occur over 12 consecutive months if that relative pace continued, placing it on a yearly-rate scale.

The result is a statement about pace: September's adjusted sales activity, if sustained for a year, would amount to 3.99 million sales. It is not the number of homes that closed in September. Nor is it simply September's raw sales multiplied by 12. The seasonal adjustment changes the monthly figure before it is annualized, so multiplying a raw monthly count by 12 would not reproduce the SAAR.

A simple comparison shows why the distinction matters.

ILLUSTRATIVE ONLY: hypothetical figures, not NAR data. Suppose a market typically records far more closings in early summer than in midwinter. A summer month and a winter month could have very different raw sales counts even if the underlying market were equally active in both. Seasonal adjustment is what makes the two months more meaningfully comparable despite their normal calendar differences. Annualization then expresses each adjusted monthly pace on the common yearly-rate scale used in NAR's Existing-Home Sales reports.

This is why NAR reports monthly sales as a SAAR, and why the headline should be read as "a 3.99 million annual pace" rather than "3.99 million homes sold."

What seasonal adjustment actually does

Home sales follow a strong calendar pattern. NAR notes that sales activity tends to increase in spring and summer and slow in winter, with January typically the slowest month. When the weather is better, more people tend to look at homes. (Source: NAR, Existing-Home Sales Explained.)

Those swings say little about whether the market is strengthening or weakening, because they recur every year. If raw monthly sales were compared directly, almost every spring would look like a surge and almost every winter like a slump.

Seasonal adjustment estimates those normal recurring patterns and removes them, so that month-to-month and quarter-to-quarter comparisons are not distorted by seasonal variation. NAR's methodology says its seasonal adjustments are determined using the X-13 variant created by the Census Bureau. (Source: NAR, Existing-Home Sales methodology.)

In practice, this means a seasonally adjusted September figure has been scaled relative to what September usually looks like. A September that is busier than a typical September will show a stronger adjusted pace, even if its raw sales are lower than those of a typical June.

Two cautions apply. Seasonal adjustment removes estimated recurring patterns, not all month-to-month noise, so adjusted figures can still move irregularly. And the adjustment factors are estimates that NAR revises periodically, a point covered later in this article.

Annualized does not mean forecast

Because the headline is expressed as an annual figure, it is sometimes read as a prediction for the year ahead. It is not.

Annualization is a scaling convention. It takes the month's seasonally adjusted pace and expresses it as the yearly total that pace would produce if it held for 12 consecutive months. (Source: NAR, Existing-Home Sales methodology.) The "if" is essential. The SAAR describes the pace observed in one month. It says nothing about whether that pace will last.

So a 3.99 million SAAR for September 2026 does not mean:

• NAR expects 3.99 million existing homes to sell over the next 12 months
• 3.99 million is the projected total for 2026
• sales are guaranteed to continue at that pace

It means only that September's relative sales pace, after seasonal adjustment, corresponds to 3.99 million sales on an annual scale. Next month's report will measure October's pace, which may be higher, lower, or the same.

Annualizing places the seasonally adjusted monthly pace on the same yearly-rate scale used for other monthly Existing-Home Sales readings and for annual-rate reporting. It does not turn one month's measurement into a forecast.

How NAR builds the Existing-Home Sales estimate

NAR's current methodology describes a sequence of steps. (Source: NAR, Existing-Home Sales methodology.)

• Sample. The monthly Existing-Home Sales indicator is based on a representative sample of 210 Boards and MLSs.
• Regional division. The raw home sales data are divided among four census regions: Northeast, South, Midwest, and West.
• Quality review. NAR economists evaluate the raw sales volume from participating Boards and MLSs to ensure accuracy. NAR lists potential problems such as changes in a board or MLS's physical jurisdiction, changes in MLS vendors or staff, lack of response, and erroneous data.
• Weighting. Once problematic data are removed, the aggregated raw volume is weighted to represent sales activity for each region. NAR says the weights are benchmarked every ten years to reflect shifts in regional demand.
• Non-seasonally adjusted volume. The weighted result is the non-seasonally adjusted volume.
• Seasonal adjustment and annualization. The non-seasonally adjusted volume is then converted into seasonally adjusted annualized rates.

1Raw monthly sales data

2Regional quality review and weighting

3Non-seasonally adjusted volume

4Seasonal adjustment

5Annualized rate

Conceptual sequence — NAR does not publish a simple arithmetic formula for this estimate.

The national figure is therefore an estimate built from a sample, weighted to represent regional activity, and then adjusted and annualized. It is not a complete count of every closing in the country, and NAR's weighting details are not reproduced here.

The report also includes other statistics, such as median sales prices and inventory. Those are separate measures produced alongside the sales estimate, not components of the 3.99 million figure.

What the latest September 2026 report shows

NAR's October 13, 2026 release reported:

Existing-home sales changed across all four regions

Region	SAAR	Change vs. August 2026	Change vs. September 2025
Northeast	470,000	-2.1%	-4.1%
Midwest	930,000	0.0%	-2.1%
South	1.86 million	+2.8%	-0.5%
West	730,000	+2.8%	-5.2%

Swipe horizontally to view all columns →

Source: NAR, Existing-Home Sales release, October 13, 2026. All sales figures are seasonally adjusted annual rates, not September's raw transaction counts.

Existing-home sales · September 2026

3.99M

SAAR

+1.5%

Month over month

-2.2%

Year over year

Seasonally adjusted annual sales pace, not September's raw transaction count.

Source: National Association of REALTORS®

Read correctly, existing-home sales increased 1.5% from August to September 2026, to a seasonally adjusted annual rate of 3.99 million, and were 2.2% lower than in September 2025. The 1.5% monthly change means the seasonally adjusted annual pace was 1.5% higher than August's. It is not a comparison of raw, unadjusted closings in the two months. The −2.2% annual change compares September 2026's sales pace with September 2025's. Neither figure is a price change.

Regionally, the monthly pace rose in the South and West, held steady in the Midwest, and declined in the Northeast, while sales were lower than a year earlier in all four regions. These figures describe the pace of completed sales, not the quality of any regional market.

The same release included supporting figures. Total housing inventory was 1.60 million units, up 1.9% from August and 4.6% from a year earlier, equal to a 4.8-month supply, unchanged from August and up from 4.5 months a year earlier. The median existing-home price was $417,500, up 1.3% from $412,300 a year earlier. (Same source.) Those are separate measures of for-sale stock and price, discussed briefly below.

NAR's interpretation should be kept apart from the measured data. NAR Chief Economist Lawrence Yun described the September gain as small on a seasonally adjusted basis, said August and September marked the two slowest months for home sales this year, and said further increases in mortgage rates mean fewer buyers can qualify. (Same source.) That is NAR's explanation. The sales figures themselves record how fast completed sales occurred, not why.

Why sales volume and median price can move differently

The Existing-Home Sales report contains both a sales figure and a price figure, and the two answer different questions.

Existing-Home Sales

Completed transaction activity

Pending Home Sales

Signed-contract activity

Housing inventory

For-sale stock

FHFA HPI

House-price change

Closings, contracts, stock, and price change are different measures.

The September figures illustrate why the two should not be compared directly. The +1.5% is a monthly change in the seasonally adjusted sales pace. The +1.3% is a year-over-year change in the median price. They measure different things over different periods, so lining them up says nothing about how sales and prices relate.

Sales volume and prices can also move in different directions. More transactions do not mean higher prices, and fewer transactions do not mean lower prices. NAR notes that the only valid comparisons for median prices are with the same period a year earlier, because of seasonality in buying patterns. (Source: NAR, October 13, 2026 release.)

A median sale price also reflects which homes happened to sell in a period, so it can shift when the mix of homes changes. For a measure designed to track price change for the same properties over time, see how the FHFA House Price Index measures price change.

Existing-home sales vs. pending-home sales

NAR publishes two related measures of resale activity at different stages of a transaction.

• Pending Home Sales track signed contracts that have not yet closed.
• Existing-Home Sales track completed transactions.

Because a contract is signed before the sale closes, NAR describes pending sales as generally leading existing-home sales by a month or two. (Source: NAR, Pending Home Sales page.)

The two can move in different directions in the same month. In September 2026, existing-home sales rose 1.5% from August on a seasonally adjusted basis, while NAR's Pending Home Sales release reported pending-contract activity down 7.3% over the same months. (Sources: NAR, Existing-Home Sales release, October 13, 2026; Pending Home Sales release, October 20, 2026.) This is not a contradiction. Many of September's closings went under contract in earlier months, while September's contracts will mostly close later, and some will not close at all.

The Insight on how pending home sales measure signed contracts before closing covers the contract side in detail.

Existing-home sales vs. housing inventory

The monthly report also states how many homes are available for sale, but inventory is a different kind of number from sales.

• Existing-home sales measure a flow: how fast completed transactions are occurring, expressed as an annual rate.
• Housing inventory measures a stock: how many homes are for sale at a point in time.

In September 2026, NAR reported a 3.99 million sales rate alongside 1.60 million units of inventory. (Source: NAR, October 13, 2026 release.) Those two numbers should not be compared as if they measured the same thing, or read as "homes sold versus homes listed." One is an annualized pace of closings. The other is a count of homes on the market. NAR's months'-supply figure relates the two, but that is a separate ratio with its own interpretation.

The Insight on what housing inventory and months of supply actually measure explains inventory and months of supply in detail.

How to compare Existing-Home Sales correctly

Existing-home sales figures are only comparable when the measurement is held constant. Before comparing two figures, check:

• The same series: total existing-home sales with total, or single-family with single-family, not a mix of categories.

• The same geography: national with national, or the same region with the same region.

• The same seasonal basis: seasonally adjusted annual rates with other seasonally adjusted annual rates, not with unadjusted monthly counts.

• The same comparison period: month over month with month over month, or year over year with year over year.

• The same release vintage, where practical: because seasonally adjusted figures can be revised.

• Rate vs. count: a SAAR is an annualized pace, not the number of homes that sold in the month.

• Sales vs. price: a change in sales volume is not a change in price.

• Existing vs. new: NAR's existing-home sales and the Census Bureau's new-home sales are separate series. NAR notes that existing-home sales are based on closings, while the Census Bureau's new single-family home sales series is based on contracts or the acceptance of a deposit, so the two can move in different directions in the same month. (Source: NAR, October 13, 2026 release.)

One error is worth naming specifically. Comparing September's 3.99 million sales rate with its 1.60 million inventory figure, as if both counted homes in the same way, mixes an annualized flow with a point-in-time stock.

For local housing context, you can explore Metro Analytics market pages. The Metro Analytics methodology explains how the measures on those pages are defined and sourced.

Why previously published numbers can be revised

A seasonally adjusted figure published one month may be different when you look it up later. That is expected, and it does not mean the original report was wrong.

NAR's methodology says it revises its seasonal adjustment factors and seasonally adjusted annual rate data each year, with the February release of revised December data. The annual revision affects seasonally adjusted national and regional data for existing-home sales, single-family homes, and condo/co-op homes, as well as months' supply of inventory. In that revision, seasonally adjusted data for the past three years are revised. (Source: NAR, Existing-Home Sales methodology.)

The reason is that seasonal factors are estimates. As more data accumulate, the estimated seasonal pattern can be refined, and earlier seasonally adjusted figures are recalculated to reflect the updated factors.

Two practical points follow. A SAAR quoted from an older release may not match the current published value for the same month. And comparisons are cleanest when both figures come from the same release vintage. Revisions are part of how the series is maintained, not a sign that a particular month's report was in error.

What Existing-Home Sales can—and cannot—tell you

Used carefully, the Existing-Home Sales report can help describe:

• The pace of completed resale transactions: how fast existing homes are changing hands, on a seasonally adjusted annual basis.
• The direction of resale activity: whether the completed-sales pace rose or fell from the prior month or year.
• Regional differences: how the sales pace compares across the four census regions.
• Change over time: when the same series, basis, and vintage are used.

On its own, the report's sales figure cannot tell you:

• the pipeline of contracts not yet closed
• whether inventory is adequate, or whether there is a structural housing shortage
• whether housing is affordable
• where home prices are heading
• buyer demand in isolation
• future mortgage rates
• future sales
• investment returns, or whether a market is "good" or "bad"

A rising sales pace is not, by itself, rising demand. Completed sales reflect buyers, but also how many homes were available, financing conditions, how many earlier contracts reached closing, and timing. A 1.5% increase in the sales pace is not a 1.5% increase in demand.

The SAAR is a measurement of observed activity. It is not a forecast of sales or prices, and it is not a recommendation about any market.

The takeaway

Existing-home sales measure completed transactions on previously owned single-family homes, condos, and co-ops. NAR's headline is a seasonally adjusted annual rate. September 2026's 3.99 million is an annualized pace, adjusted for normal seasonal patterns, not the number of homes that sold in September and not a forecast of the next 12 months.

Seasonal adjustment and annualization do different jobs: one removes the normal calendar swing, the other puts the adjusted pace on a yearly scale. Reading the report well also means keeping its measures separate. Sales volume is not price. Sales volume is not inventory. And completed existing-home sales are not pending contracts, which come earlier in the transaction.

To see local housing context for a specific area, explore Metro Analytics Markets, or view the Methodology to see how each measure is built.