Have Housing Starts Peaked In 2018?


The Census Bureau report released today shows that housing starts roughly came in line with estimates, coming in at 1,201,000. Additionally, the previous month’s numbers were revised slightly lower.  Even though some people are disappointed with the housing starts this year, the numbers look perfectly right to me because they show a slow tortoise crawl, as I expected. These numbers should be taken in the context of higher labor cost, higher mortgage rates and increases in existing inventory, which provide a less expensive home compared to new home.

October CENUSUS Housing Starts

From Census: https://www.census.gov/construction/nrc/pdf/newresconst.pdf

Building Permits:

Privately‐owned housing units authorized by building permits in September were at a seasonally adjusted annual rate of 1,241,000. This is 0.6 percent (±1.2 percent)* below the revised August rate of 1,249,000 and is 1.0 percent
(±1.2 percent)* below the September 2017 rate of 1,254,000. Single‐family authorizations in September were at a rate of 851,000; this is 2.9 percent (±1.2 percent) above the revised August figure of 827,000. Authorizations of units in buildings with five units or more were at a rate of 351,000 in September.

Housing Starts:

Privately‐owned housing starts in September were at a seasonally adjusted annual rate of 1,201,000. This is 5.3 percent (±11.3 percent)* below the revised August estimate of 1,268,000, but is 3.7 percent (±12.1 percent)* above the September 2017 rate of 1,158,000. Single‐family housing starts in September were at a rate of 871,000; this is 0.9 percent (±8.9 percent)* below the revised August figure of 879,000. The September rate for units in buildings with five units or more was 324,000.  

When I am asked what would let me know if housing starts have peaked, “I tell people if new home sales start to show negative year over year sales and monthly supply stays over 6.5 months consistently, then we can start to be concerned– but that has not happened yet”.  Next year if more, cheaper existing inventory comes to the market things could get interesting. This could create more slowing in housing starts in 2019. Housing starts like new home sales can be very volatile month to month, which is why revisions and trends are key.

From Doug Short:

October ADJ Housing Starts

In order for total housing starts to grow, we will need more growth in single family starts. For acceleration in single family starts we need more new home sales. Today, new home sales are trending around 639,000 and I believe that is where they should be. I don’t subscribe to idea that new home sales are so strong that they warrant a boom in housing construction. In fact my call that housing starts won’t reach 1,500,000 this decade looks really good right now. The modest growth in new home sales need to be considered in the context of the current economy. We are in the longest job expansion ever in U.S. history, soon to be the longest economic expansion ever, with mortgage rates below 5% since early 2011. Still for 2018, new home sales are beating my sales expectation.  The median sales price is cooling off and this is a plus for demand. New home sales are growing as builders include smaller homes in the mix.  It will be interesting to see if rising existing inventory impacts sales in certain markets like coastal hot spots. These areas are seeing an increase in inventory year over year and declining sales, year over year. This is happening in California and especially in the Seattle marketplace.

From Calculated Risk:

Single family starts are up 6.0% year to date which is a good print.  Total starts are up 6.4% year to date. Yes, total starts are still up year to date. This growth looks normal to me considering how new home sales have been trending in this cycle. The slow and steady housing start story can continue as long as new home sales grow. The number of construction job openings are at cycle highs with 298,000 construction jobs openings and over 7,200,000 construction workers. Total employment for construction is roughly 450,000 jobs away from what it was during the peak of the housing bubble when housing starts were 1,000,000 units more than where they are at today. Even if we just count residential workers, we are not that far from the total employment levels we saw at the peak of the housing bubble.


October JOMTS construction

From Fred: 
October Construction total employment

From Fred:
October residential construction employment

While this hasn’t been the best few months for housing starts, keep in mind that the trend matters more than any one report, either positive and negative. The housing story has remained the same for years now; slow and steady growth for new home sales and housing starts. Still, housing starts in any context are still low.

One final note for those in Southern California: I will be part of a housing panel on November 8th in Mission Viejo to talk about the state of housing and what to look for in 2019.


Logan Mohtashami is a financial writer and blogger covering the U.S. economy with a specialization in the housing market. Logan Mohtashami is a senior loan officer at AMC Lending Group,  which has been providing mortgage services for California residents since 1987. Logan also tracks all economic data  daily on his own facebook page https://www.facebook.com/Logan.Mohtashami