Importance
Housing starts reflect the commitment of builders to new construction activity. Purchases of household furnishings and appliances quickly follow.

Interpretation
The bond market will rally when housing starts decrease, but bond prices will fall when housing starts post healthy gains. A strong housing market is bullish for the stock market because the ripple effect of housing to consumer durable purchases spurs corporate profits. In turn, low interest rates encourage housing construction.

The level as well as changes in housing starts reveal residential construction trends. Housing starts are subject to substantial monthly volatility, especially during winter months. Following either the three-month moving averages of the monthly levels or year-over-year percent changes will smooth out monthly volatility and provide a more accurate picture of the underlying trend.

It is useful to examine the trends in construction activity for single homes and multi-family units separately because they can deviate significantly. Single-family home-building is larger and less volatile than multi-family construction. It is more sensitive to interest rate changes and less speculative in nature. The construction of multi-family units can be substantially influenced by changes in the tax code and speculative real estate investors.

Housing construction varies by region as well. The regions of the United States don't all follow exactly the same economic patterns because industry concentration varies in the four major regions of the country. The regional dispersion can mask underlying trends. It is important to look at the total level of housing construction but also at the regional distribution.

Frequency
Monthly.

Source
Bureau of the Census, U.S. Department of Commerce.

Availability
Usually during the third week of the month

Coverarge
Data are for the previous month. (Data for June are released in July.)

Revisions
Monthly, data for the prior two months are revised to incorporate more complete information. Annually, new seasonal adjustment factors are introduced in February with the release of the January data. These revisions affect at least three years of data. The magnitude of the revisions is typically small.