Housing Starts Sink to 6-Year Low

Background on Housing Starts

Housing starts track how many residential buildings began construction in the preceding month. The data is divided into three types of structures: single-family homes, residences with 2-4 units (condos or townhouses), and structures with 5+ units (apartment complexes). A critical aspect of the home-building industry is the powerful influence it has on the rest of the economy.

Here is the historical series for total privately-owned housing starts, which dates from 1959. Because of the extreme volatility of the monthly data points, a six-month moving average has been included.

Housing Starts: Structure Breakdown

In May, single-family housing starts were at a seasonally adjusted annual rate of 0.882 million. This represents a 1.9% decrease from April and a 6.7% decline from the previous year.

Multi-family buildings housing starts were at a seasonally adjusted annual rate of 0.284 million. This represents a 41.6% decrease from April and a 12.3% decline from the previous year.

Note: buildings with 2-4 units only have N.S.A. data available because tests for identifiable and stable seasonality do not meet reliability standards.

Housing Starts: The Population-Adjusted Reality

The Census Bureau's mid-month population estimates show substantial growth in the US population since 1959. Here is the data with a simple population adjustment showing housing starts as a percent of the population with a linear regression to highlight the trend. In May, housing starts as a percent of the population was at 0.34%, the lowest level in six years.

Housing Starts: A Footnote on Volatility

The extreme volatility of this monthly indicator is the rationale for paying more attention to its six-month moving average than to its noisy monthly change. Over the complete data series, the absolute month-over-month average percent change was 6.3%. The month-over-month range minimum was -26.4% and the maximum was 29.3%.

For visual confirmation of the volatility, here is a snapshot of the monthly percent change since 1960.

A Long-Term Look: Residential Building Permits and Housing Starts

In addition to housing starts, the U.S. Census Bureau and the Department of Housing and Urban Development published their findings for new residential building permits. Building permits typically preceded housing starts, offering a glimpse into future construction activity. Despite the fact that both are monthly seasonally adjusted annualized rate (SAAR) series, they are exceptionally volatile and subject to extensive revisions. Thus it is unwise to assign much significance to a single month.

Over the long haul, however, the two offer a compelling study of trends in residential real estate, especially when we adjust the permits and starts for population growth. Here is an overlay of the two series since the 1959 inception of the starts data and the 1960 inception of the permits data. The monthly data points are preserved as faint dots. The trends are illustrated with six-month moving averages of data divided by the Census Bureau's mid-month population estimates.

Here is a closer look at the overlay since 1990.

ETFs associated with home builders include: Invesco Dynamic Building & Construction ETF (PKB), iShares U.S. Home Construction ETF (ITB) and SPDR S&P Homebuilders ETF (XHB).

ETFs associated with residential real estate include: iShares Residential and Multisector Real Estate ETF (REZ).

Read more updates by Jen Nash