by Calculated Risk on 8/09/2009 02:03:00 PM

Professor Arthur C. Nelson, Director of the Metropolitan Research Center at the University of Utah, has kindly sent me his new paper: "The New Urbanity: The Rise of a New America" (no link). Nelson sees a dramatic shift in American cities based on changes in demographics and in housing peferences. He believes this will lead to a "new era of infill and redevelopment."

Nelson also argues this will lead to a decline in the homeownership rate.

Note: Brief excerpts of Dr. Nelson's paper removed by request.

This graph shows the homeownership by age group for three different time periods: 1985, 2000, and 2007. Back in 1985, the homeownership rate declined significantly after people turned 70. However, more recently, the homeownership rate has stayed above 80% for those in the 70 to 75 cohort, and close to 80% for people over 75.

I noted in April:

I expect the homeownership rate to remain high for the boomer generation too. Although there will probably be a geographic shift as the boomer generation retires (towards the sun states) and some downsizing, I don't think the aging of the boomer generation will negatively impact the homeownership rate for 15 years or more.

So Dr. Nelson is coming to a different conclusion. He thinks the homeownership rate will fall to 63.5% by 2020, and I think it will stay a little higher as many older couples and singles stay in their homes.

This has significant implications for planning and homebuilders. If Nelson is correct, there will be a dramatic shift towards a "new urbanity" and away from suburbs. And also a shift towards more renting.

Professor Nelson sent me this projection:

Home Tenure	2005	2020	Change	Percent Change in Supply	Share of Change in Supply	Tenure 2005	Tenure 2020
Owner-Homes (occupied and vacant)	74,164	87,135	12,971	17%	43%	68.90%	63.50%
Renter-Homes (occupied and vacant)	35,903	53,254	17,351	48%	57%	31.10%	36.50%
Total	110,067	140,389	30,322
Source: Arthur C. Nelson, Director, Metropolitan Research Center

Notice the shift to rental units in Nelson's projections. Nelson does note that there were over 2 million excess units in 2005, and that needs to worked off first. Nelson is projecting a need for 30 million new housing units over the next 15 years (I think this is high, but Nelson is expecting many more single person households).

This has huge implications for builders. Using Nelson's figures, home builders will only have to build about 800 thousand (on average) single family units per year through 2020 (after the excess is worked off). This is far below the 1.25 million per year seen in 2004 and 2005. That level of production is not coming back. Here is something I wrote in 2007: Home Builders and Homeownership Rates

[With the rising homeownership rate] the homebuilders was ... had the wind to their backs. Instead of 800K of new owner demand per year (plus replacement of demolished units, and second home buying), the homebuilders saw an additional 500K of new owner demand during the period 1995 to 2005. This doesn't include the extra demand from speculative buying. Some of this demand was satisfied by condo conversions and owner built units ...

Looking ahead, if the homeownership rate stays steady, the demand for net additional homeowner occupied units would fall back to 800K or so per year (assuming steady population growth and persons per household). However the homeownership rate is declining, and this is now a headwind for the builders.

It appears the rate is declining at about 0.33% per year (Goldman's Hatzius estimated 0.5% per year). This would mean the net demand for owner occupied units would be 833K minus about 333K or 500K per year - about 40% of the net demand for owner occupied units for the period 1995 to 2005.

This means the builders have two problems over the next few years: 1) too much inventory, and 2) demand will be significantly lower over the next few years than the 1995-2005 period, and even when the homeownership rate stabilizes and the inventory is reduced, demand (excluding speculation) will only be about 2/3 of the 1995-2005 period.

I think Nelson is correct about the trend, but might be overestimating the shift towards renting. Also I think using age 75 would be better for figure 1, so I think this shift will be delayed by about 10 years.