New Residential Construction (Housing Starts and Permits)

What the report measures

The New Residential Construction report, usually called housing starts, tracks the early stages of building new homes in the United States. It is produced jointly by the U.S. Census Bureau and the Department of Housing and Urban Development (HUD) and covers new, privately owned housing units, excluding manufactured (mobile) homes.

It reports three main stages:

• Building permits: units authorized by local permit-issuing offices. A permit is the step before construction.
• Housing starts: units on which construction has begun, defined as the start of excavation for the foundation.
• Housing completions: units finished during the month.

It also reports units authorized but not yet started and units under construction, which together show the pipeline of homes on the way.

Who publishes it and when

The report comes out monthly at 8:30 AM Eastern Time, typically in the second half of the month after the month it covers. Release information and the data series are on the Census Bureau New Residential Construction page.

The numbers traders watch

The headline figures are expressed as a seasonally adjusted annual rate (SAAR): the number of units that would be started over a full year if that month’s pace continued. This makes months comparable with each other and with annual totals, but it means the headline is not the number of homes actually started in the month.

Within the totals, the most watched details are:

• Single-family starts and permits. Single-family construction is the steadier part of the market and the part most directly tied to household demand and mortgage rates.
• Buildings with five units or more. Apartment construction is lumpy, because one large project can swing the monthly number. Strong or weak multifamily figures often account for big headline moves.
• Permits as a leading signal. Because permits come before starts, many analysts treat them as the better guide to where construction is heading.
• Regional breakdowns, for the Northeast, Midwest, South and West.

Why monthly moves are noisy

Housing starts are estimated from a sample survey, and the sampling error is large. The release attaches a 90 percent confidence interval to each monthly change in starts and completions, and those intervals are often wider than the reported change itself. When that happens, the release marks the change with an asterisk to show there is not enough statistical evidence to conclude that the actual change differs from zero.

Weather adds more noise. A cold or wet month can delay starts, and a mild one can pull them forward, even after seasonal adjustment. For both reasons, analysts usually focus on three-month averages and on the trend in permits rather than any single month.

Revisions

Each release revises the previous month’s figures, and earlier months can be revised as late reports arrive. Seasonal factors are updated periodically, which can change the history of the series. The first estimate of any month should be treated as provisional.

How markets commonly react

Housing starts generally move the broad stock and bond markets less than reports such as the jobs report or the CPI. The more noticeable reactions tend to be in the shares of homebuilders, building materials suppliers and related industries. Strong construction is commonly read as a sign of economic momentum and, potentially, of demand that could keep interest rates higher; weak construction can be read as a sign that higher mortgage rates are biting. Given the noise in the data, a single month’s surprise is often discounted, and any reaction is unreliable.

Why housing matters to the economy

Residential investment is a relatively small share of GDP, but it is sensitive to interest rates and tends to turn ahead of the wider economy. New construction also drives demand for materials, appliances and furnishings, and employs builders and tradespeople. That is why economists watch it as an early read on how monetary policy is affecting activity. Existing home sales, which are not part of this report, are a separate indicator.

Common misreadings

• Reading the SAAR as a monthly count. The headline is an annualised pace.
• Over-reading one month. Confidence intervals are wide, and many monthly changes are not statistically significant.
• Ignoring the multifamily swing. A headline jump driven by apartments says less about home buyers than one driven by single-family homes.

Frequently asked questions

What time are housing starts released?

At 8:30 AM Eastern Time on the scheduled release date.

What is the difference between building permits and housing starts?

A permit is an authorization to build issued by a local government. A start is recorded when construction actually begins. Permits usually lead starts.

What does “seasonally adjusted annual rate” mean?

It is the number of units that would be started in a year if the month’s seasonally adjusted pace continued for twelve months.

Why do housing starts jump around so much?

The survey has large sampling error, apartment projects are lumpy, and weather can shift construction from one month to another.

Explainer updated November 13, 2026.