New home sales estimate contracts and deposits for new single-family houses; the report also covers inventory, construction stage, prices, and months' supply.
New home sales estimate the number of newly built, privately owned single-family houses for which a buyer signs a sales contract or makes a deposit. In the U.S. Census Bureau’s New Residential Sales data, a house can be counted as sold before construction starts, while it is under construction, or after it is completed. The statistic does not wait for the transaction to close.
The monthly report, published jointly by the U.S. Census Bureau and the Department of Housing and Urban Development (HUD), also estimates new houses for sale, months’ supply, sales by construction stage and region, and median and average sales prices. It is useful for analyzing the new-home market, but it is a sample-based national and regional indicator, not a complete transaction registry or a valuation of every new house.
The Census Bureau defines the sale date as the date a deposit is made or a sales agreement is signed. A verbal agreement or an unfinished negotiation is not enough. A deposit can include earnest money or another good-faith payment toward the purchase.
This definition creates three important consequences.
The Survey of Construction does not follow the transaction through closing for purposes of the sales count. If a contract is later canceled or fails to close, the house remains counted as sold under the survey definition.
New home sales therefore measure commitments made during the period, not completed title transfers. Lenders, builders, and analysts should not equate the release with funded mortgages, recognized builder revenue, occupied homes, or final legal closings.
A qualifying house can be sold from plans or a model before ground is broken. It can also be sold while under construction or after completion. The report separates sold houses and for-sale inventory by these stages:
| Construction stage | Meaning | Analytical use |
|---|---|---|
| Not started | A permit has been issued, but construction has not begun | Shows demand accepted against future construction |
| Under construction | Ground has been broken, but the house is not complete | Connects sales and inventory with active building work |
| Completed | The house meets the survey’s completion definition | Identifies finished units sold or still available |
Because sales can precede starts, New Home Sales is not simply a lagged version of Housing Starts. The two series observe different events.
The sales estimate covers new single-family houses built for sale where the transaction is intended to include both the house and the land. The Census definitions exclude:
These exclusions mean the series is narrower than all new single-family construction and much narrower than all residential construction.
The survey’s single-family category includes detached houses and qualifying attached units. A side-by-side townhouse or row house can be classified as single-family when each unit has the required ground-to-roof separation, no unit above or below, separate heating, and individual utility metering under the Census definition.
Units stacked above or below one another are classified as multifamily. A condominium in a multifamily structure is therefore not included in New Residential Sales merely because the unit can be individually owned.
Physical structure, not the marketing label alone, determines classification.
New home sales are estimated through the Census Bureau’s Survey of Construction (SOC), which covers permit-issuing and non-permit areas.
In permit areas, field representatives select a sample of permits for new housing and contact owners or builders as needed. Selected houses are followed through construction and, for qualifying single-family houses built for sale, until they are sold. In sampled areas that do not require permits, field representatives canvass for new construction and follow identified projects.
The estimate is not simply a count of permits marked “sold.” Census uses permit totals, survey observations, weighting, and adjustments to produce national and regional estimates. Because a contract can be signed before a permit is issued, the methodology includes an adjustment for estimated pre-permit sales and late reports.
Monthly sales estimates are available for the United States and four Census regions. The report does not provide a statistically equivalent monthly new-home-sales estimate for every state, metropolitan area, or neighborhood.
Regional estimates usually have more sampling variability than the national total because they use fewer observations. A national or regional result should not be treated as direct evidence of conditions in a specific development.
The current New Residential Sales release includes several related measures.
| Release item | What it measures | What it does not establish |
|---|---|---|
| New houses sold | Contracts signed or deposits accepted for qualifying new single-family houses | Closings, mortgage fundings, occupancy, or builder revenue |
| New houses for sale | Qualifying houses offered for sale with no contract or accepted deposit | All vacant homes or all unsold builder lots |
| Months’ supply | For-sale inventory relative to the current sales rate | A guaranteed selling period for each house |
| Median sales price | Estimated midpoint of prices among new houses sold | Price change for a constant-quality house |
| Average sales price | Estimated arithmetic mean price | The price of a typical house when the distribution is skewed |
| Stage of construction | Sold and for-sale houses that are not started, under construction, or completed | Exact completion dates or cancellation risk |
| Regional sales | Estimated sales by Census region | Local development-level demand |
The report also provides not-seasonally-adjusted tables and additional quarterly detail, including selected sales-price and financing categories.
The headline New Home Sales number is generally a seasonally adjusted annual rate (SAAR). Seasonal adjustment attempts to remove recurring calendar patterns. Annualization multiplies the seasonally adjusted monthly value by 12.
If the seasonally adjusted estimate for a month were 60,000 houses, the annualized rate would be:
1Illustrative SAAR = 60,000 x 12 = 720,000 houses
The 720,000 figure would describe that month’s annualized sales pace. It would not mean 720,000 houses were sold during the month, and it would not predict the final total for the year.
When comparing figures, check whether each is:
Dividing SAAR by 12 recovers the seasonally adjusted monthly value used in annualization, not necessarily the published not-seasonally-adjusted monthly estimate.
A new house enters the for-sale inventory when it is being built for sale and either a permit has been issued in a permit area or foundation work has begun in a non-permit area, provided no contract has been signed and no deposit accepted.
For-sale inventory can include houses that are:
This stage mix affects interpretation. A large not-started inventory represents a different delivery and cost profile from a large completed inventory. Completed homes can be available sooner but may expose a builder to carrying costs, maintenance, incentives, and price concessions while they remain unsold.
Months’ supply relates the estimated inventory of new houses for sale to the current sales pace. When sales are expressed as SAAR, the basic relationship is:
1Months' supply = Houses for sale / (Annualized sales rate / 12)
Suppose the estimated inventory is 420,000 houses and the annualized sales rate is 720,000:
1Monthly sales pace = 720,000 / 12 = 60,000
2Months' supply = 420,000 / 60,000 = 7.0 months
This is a rate-based measure. It says the current inventory would last about seven months if the current sales pace continued and no additional houses entered inventory. Neither condition is likely to hold exactly.
Months’ supply increases when inventory rises, the sales pace falls, or both. It can fall when sales accelerate, inventory declines, or both. Always inspect the numerator and denominator rather than treating the ratio as a standalone fact.
A rise driven by more completed inventory can create different builder risks from a rise driven mainly by not-started units. Stage-of-construction tables provide that context.
The report publishes estimated median and average prices for new houses sold. The median is the midpoint of the estimated price distribution; the average is the arithmetic mean.
These price measures are not constant-quality house-price indexes. Changes can reflect both:
For example, if a larger share of sales shifts to smaller houses or lower-cost regions, the national median can fall even when comparable house prices are unchanged. Conversely, more high-priced sales can lift the average and median without broad appreciation.
Use a House Price Index when the question is quality-adjusted price movement. Use New Residential Sales price data when the question concerns the price distribution of new houses sold under this survey’s coverage.
New home sales are sample estimates, not complete counts. The release reports estimated sampling uncertainty using relative standard errors and 90% confidence intervals for percentage changes.
If a release reports a change of 4% with a margin of error of plus or minus 8%, the interval extends from -4% to +12%. Because that interval includes zero, the survey does not provide sufficient statistical evidence to conclude that the underlying direction was positive.
This does not mean the estimate is useless. It means the correct description is cautious: the point estimate increased, but the change was not statistically distinguishable from zero at the stated confidence level.
Confidence intervals address sampling variability. They do not capture every nonsampling risk, such as nonresponse, undercoverage, reporting error, processing error, or imputation.
Breaking the sample into smaller groups reduces the number of observations supporting each estimate. Regional changes and narrow price bands can therefore have larger relative standard errors than the national total. Large percentage moves in a small category may still carry substantial uncertainty.
Preliminary estimates are revised as later reports replace imputed data and more information becomes available. Under the current methodology:
The latest historical series may therefore differ from the figures available to an analyst at an earlier decision date. For backtesting or investment research, preserve the original release or identify the data vintage used.
Current Census explanatory notes also warn that seasonally adjusted sales can move irregularly and that several months are needed to establish a trend. One preliminary headline should not outweigh its confidence interval, revisions, and surrounding months.
Remember that the series records contracts and deposits, not closings. It can lead physical completion and mortgage funding, but cancellations can prevent an initial commitment from becoming a completed transaction.
Compare each monthly and year-over-year percentage change with its margin of error. Avoid declaring a confirmed acceleration or decline when the interval includes zero.
Review changes to the prior three months. A favorable current estimate can accompany downward revisions, while a weak point estimate can accompany stronger revised history.
These measures answer different questions. Sales describe new commitments, inventory describes available houses, and months’ supply scales inventory by the current selling pace.
Determine whether sold and for-sale houses are not started, under construction, or completed. The stage mix affects delivery timing, construction exposure, and carrying costs.
National movement may conceal offsetting regional results. Treat volatile regional estimates carefully and supplement them with local permits, listings, builder data, and transaction evidence.
Use Housing Starts for units entering construction, Existing Home Sales for completed transactions in the much larger existing-home market, and the Weekly Mortgage Applications Survey for participating lenders’ mortgage application activity.
Assume an illustrative monthly release shows:
| Measure | Prior month | Current month | Change |
|---|---|---|---|
| Sales rate (SAAR) | 780,000 | 720,000 | -7.7% |
| Houses for sale | 400,000 | 420,000 | +5.0% |
| Months’ supply | 6.2 | 7.0 | +0.8 month |
| Completed houses for sale | 90,000 | 105,000 | +16.7% |
The ratio calculation for the current month is:
1420,000 / (720,000 / 12) = 7.0 months
The initial interpretation is that sales commitments slowed while inventory grew, increasing supply relative to the sales pace. The increase in completed houses for sale could also indicate more finished inventory awaiting buyers.
That conclusion remains incomplete until the analyst checks:
The example does not prove that builders will cut prices or that all local markets are oversupplied. It identifies evidence that should prompt deeper review.
| Indicator | Event measured | Coverage emphasis | Timing distinction |
|---|---|---|---|
| New Home Sales | Contract signed or deposit accepted | Qualifying new single-family houses built for sale | Can occur before construction starts and before closing |
| Housing Starts | Excavation begins for footings or foundation | New privately owned housing units, including single-family and multifamily | Physical construction begins |
| Existing Home Sales | Transaction closes for an existing property under the publisher’s methodology | Previously owned single-family homes and other covered existing properties | Reflects completed transactions rather than new-home contracts |
| Mortgage applications | A mortgage application enters a participating lender’s survey data | Purchase and refinance applications under survey rules | Application can precede approval, funding, and closing |
No one series is a substitute for the others. Together they show different parts of the housing and financing process.
The report provides market context for public and private builders, but national sales are not the same as a builder’s net orders, backlog, deliveries, closings, revenue, or gross margin. Company disclosures may treat cancellations, joint ventures, incentives, and geographic mix differently.
Useful comparisons include:
A signed contract can create future purchase-mortgage demand, but application, underwriting, approval, rate lock, funding, and closing remain separate events. Mortgage rates, borrower capacity, appraisal results, and completion delays can change the outcome.
Lenders financing builders should also distinguish sold-not-started houses, sold houses under construction, completed unsold inventory, and land or lots that are not represented in the sales count.
Sales commitments can influence builders’ decisions to start additional units, but current construction activity depends on existing backlog, labor, materials, permits, financing, and inventory. The report should be compared with permits, starts, construction spending, and supplier order data.
New home sales can provide timely evidence about demand for newly built single-family housing. It does not by itself measure household confidence, economic growth, recession risk, affordability, or total housing turnover. Those conclusions require broader evidence.
Potential influences include:
These factors can interact. The national release cannot identify the cause of a monthly movement by itself.
Before using New Home Sales data, verify:
New Home Sales data are educational market evidence, not personalized investment, lending, appraisal, legal, or tax advice. Verify the latest Census/HUD release and use local, company-specific, and transaction-specific information for financial decisions.