When is the next housing starts report?

What it measures

This release counts building permits issued, homes on which construction began, and homes completed. Because building is financed with borrowed money, it responds to interest rates sooner than almost anything else, which makes it an early read on whether monetary policy is landing.

What it does not say

Starts are not a measure of housing supply and they are not stable. The monthly figure carries a wide margin of error — often wide enough that the Census Bureau cannot say whether the change had a direction at all, a caveat printed in the release and dropped from most coverage of it. Permits are the steadier series, and a single month of either is weather as much as it is economics.

Commonly confused: starts and permits

Permits are issued before construction begins; starts are counted when ground is broken; completions come a year or more later. The three describe the same pipeline at different points, and a month where permits rise and starts fall is usually timing rather than a turn. Permits carry the smaller margin of error.

Where to look first when it lands

• Permits rather than starts, for a signal with less noise in it.
• Single-family against multi-family, which respond to different pressures.
• Whether the release states its own margin of error covers zero, which it often does.

No forecast, on purpose

There is no consensus figure on this page. Consensus is a licensed product — a survey of economists, compiled and sold by a data vendor — and we do not license it. A date with no forecast beside it is a fact. A date with an invented forecast beside it is a fabrication with a decimal point on it.