A comment was posted to the blog from September 27, MIT Prof Forecasts Quick Return to Housing Construction, requesting if this analysis was available for Houston, Texas. Rather answer in that posting, here is a new post addressing this topic for those wishing to replicate this type of analysis in their respective markets, albeit in a slightly different manner.
Most states have either a state data center or state demographer that forecasts population growth by city, municipality, metropolitan statistical area (MSA), or county (or the equivalent). The US Census Bureau has a State Data Center listing. Click here
It may take some searching on these state-specific sites, but at least in 40 of these states you will eventually find population forecasts. If the population forecast data are not available there, then search one or more search engines for state population forecasts or the state demographer. Note: These forecasts are often done under several different scenarios based on differing historical patterns. Select the one you believe is most appropriate for the locale.
Unfortunately, the detailed home occupancy and vacancy data as of Q2 2010 do not exist below the federal level. So here is a proxy method to get there.
Year-Round Vacancy Rates (by state as a proxy with 2009 as the most current)
First, just take the state year-round vacancy rate as a proxy and contrast to the population growth rate. Denver-Aurora, Colorado, for example, shows a year-round vacancy rate of 9.8 percent (proxy from the state rate). Assuming 7 percent is normal (but you pick here) you can get an idea as to how long things will remain down. A combination of job growth and eventually population growth will absorb the excess vacancy (and for now we are simplifying demolitions by ignoring them). Denver-Aurora’s projected MSA population growth rate is projected to be 46,100 per year in the next 10 years. (Remember and we will get back to this number at the bottom).
Using employment data, calculate the total change in employment in the local community from the end of 1999 to the end of 2009 (the latest 10-years of data) and contrast that to the total construction in that same period. Fortunately, the Real Estate Center at Texas A&M University has both of these data series at one site:
Residential Building Permits
Alternatively, greater resolution data (for all of the data above) are available from the U.S. Census Bureau and the Bureau of Labor Statistics:
Metropolitan Statistical Area(MSA) and States
MSAs, Counties Etc
Using Denver, for example.
So with the 46,100 population growth forecast per year and the 91,670 surplus units, it will be at least two more years before Denver’s ‘normal’ housing construction rate returns.
If you have a better methodology to handle this—then post a comment and we can all discuss.