Let’s start with a quick history lesson. From 1978 until the end of 2007, the annual housing starts nationwide (single-family and multifamily units) averaged about 1,544,000 annually. In 2008, housing starts dropped below a million to about 906,000. And in 2009, that number slid even further, to just about 550,000.
But when you factor in the robust years of 2003 to the end of 2007, housing production rose above the historical average to about 1.8 million units a year, a difference of about 256,000 units a year from the historical average. Multiplied out over the 2003-2007 cycle, this means that about 1,280,000 more units were produced than the historical average.
If we factor in the 2008 and 2009 numbers against the historical average, the number of units produced would seemingly correct the overproduction during the robust cycle of 2003-2007. That would seem to indicate that we are already in the trough of unbuilt demand and that, in 2010, we should see an increase in housing demand, with returns to historical averages sometime later this year.
But housing starts are still fumbling about, especially in some markets. What could cause this discrepancy?
Demographics may hold part of the answer. Policymakers predicate growth-management decisions (whether to hold growth boundaries tight, increase density, etc.) on population growth, the numbers of individuals per household, and a variety of other data. But sometimes the data contains assumptions. For instance, if the Portland/ Vancouver metropolitan area has historically grown at a certain percentage, an assumption can be made that this rate will continue into the future. Likewise, if the number of individuals per household shrinks because families are getting smaller, we assume that more households will be needed to accommodate population growth. Demographers label this the “headship rate.”
Population growth and headship rate appear to be the biggest demographic determiners of housing demand.
Over the past 25 years, internal U.S. population growth remained stable, trending neither upward nor downward.
At the tail end of the baby boom in the 1960s, births in the United States were about 4.3 million annually. By the 1970s, births had declined to about 3.2 million before rebounding to about 3.9 million annually by the late 1980s. And the number of people dying each year has gradually risen, thus contributing to the slowdown in internal population growth. But internal population growth is only part of the equation: growing immigration numbers are also critical to understanding the complete picture.
Immigration into the U.S. has offset the decline in the natural increase of the population. Net migration into the U.S. increased from about 500,000 annually in the 1980s to about 1.3 million annually in the 1990s. The last decade saw immigration increase to about 1.4 million in 2001 and then decrease to about 1 million people a year in 2007, possibly as a result of the country’s high unemployment. Current data suggests immigration numbers increasing this year through 2014.
Combining the natural population growth with immigration, we see that the U.S. added about 1.35 million households per year from 2000 to 2006, and 1.3 million in 2007. The Social Security Administration estimates that the figure will rise to about 1.36 million from 2010 to 2012 in total growth. But the headship rate may either supplement or restrain population growth.
Understanding the headship rate is important to ascertaining where things are headed. Every household consists of individuals who occupy the same unit. Sometimes it is an individual, a family, or a group of unrelated individuals. But every household contains only one head for statistical purposes. In order to ascertain the headship rate, one must examine how many heads of household there are within an age group as compared to the number of individuals within that age group.
Throughout the 1960s to the late 1970s, headship rates rose, reflecting the social changes going on in the U.S. Divorce rates increased, individuals began marrying later, and the entrance of women into the workforce and women becoming heads of their own households all contributed to an increase in the headship rates, but decreased the number of individuals within the household. Since the 1990s, headship rates have declined, partly as a result of changes in the elderly population whereby both spouses survive into old age. Senior citizens, if married, continue for statistical reasons to be counted as having only one household even though one spouse may be in assisted living and one spouse is at home.Thus headship rates decrease because both spouses remain living.
But that still isn’t the end of the necessary data points. The other factors include scrappage and the number of vacant units in the marketplace. “Scrappage” refers to the number of units lost to natural disasters, dilapidation, and conversion to nonresidential use. Examining scrappage yields no discernible cycle. Some strong housing-demand years, such as 1979 and 2000, had high removal rates, as well as low-demand years, such as 1981 and 1990. And so most demographers apply an average developed over the preceding ten years, which currently equates to the need to replace about 250,000 scrapped units a year.
And finally we get to the issue of vacant properties. Vacant properties fall into two categories. Utilized properties are properties used for vacation, seasonal properties, properties for migratory workers, those sold awaiting occupancy, or those held off the market by the owner for some other reason, such as dealing with an estate. Unutilized properties typically are those that are listed for sale or listed for rent.
From 1965 to about 2000, utilized vacant homes as a total percentage of the housing stock floated between 7 and 8 percent, with declines in percentages beginning about the time a recession began and often continuing for a couple of years before rebounding. In 2000, the percentage of utilized housing units began to rise dramatically, almost reaching 10 percent of the total housing stock before the current recession.
This suggests what we anecdotally know: that investors began to shift into real estate after the technology bubble burst in the late 1990s. But just as many primary residences are in financial trouble, factoring in second homes with financial issues exacerbates the supply, with utilized vacant properties being sold or foreclosed in high numbers.
Census data suggests that nationally there is still surplus inventory of about 1.6 million units. But this surplus inventory doesn’t take into account the current delay or failure of new households to come to the market.
A good example is recent college graduates who return home to reside with their parents rather than joining the workforce and acquiring their own housing, either rented or owned. Another example is a phenomenon that we are seeing in the Portland/Vancouver metropolitan market: extensive group-shared housing. Our market is seeing high migration to the Northwest despite high unemployment numbers. Individuals are reducing costs by entering into shared housing leases.
Anecdotally we are told by property managers that multiparty leases for apartments have doubled in the last two years. But once the job market improves, these individuals will create their own households, thus increasing the headship rate and reducing the supply.
So what does this collection of national data mean for the Northwest? While our Northwest real estate market isn’t as saturated with product as some other markets have been because of growth restrictions, we still face challenges in that our unemployment rate remains high. Furthermore, current housing data suggests that most sales occurred as a result of government intervention into the housing market by keeping interest rates at historical lows and by making the recent tax credits available for first-time purchasers and now for all purchasers. And adding to the complexity are the looming potential issues with commercial real estate.
Despite these issues, for the Northwest markets, I would expect to see the homebuilding industry begin to recover starting in the middle of second quarter 2010, with strong numbers by the fourth quarter of this year. And if the job situation improves later in 2010 and the new heads of households emerge, then we should see a sustained improvement back to historical norms—think 2002, not 2005.
And while I am an eternal optimist, I know there are some risks to a stable sustained housing recovery. I do not pretend to understand all the complex rules regulating the banking industry or other lending institutions. Pressures on those industries may force new building or new development to the sidelines until the employment picture improves. But what I do know is that the data above suggests that if the industry does not see an increase in production in the total number of units, whether single-family or multifamily, we will be headed for a supply shortage in the next couple of years.
My strongest fear is that by sitting on the sidelines until banks loosen up lending, we might be headed for a whipsaw in the real estate market and cascade right back into frenzied times when the supply cannot meet the demand, forcing prices to unsustainable levels.
There are solutions to assist the banks in making prudent lending decisions. One is for the home building industry to focus on where demand appears to be the heaviest; right now, that is new entry-level construction.
We should see the home builders modify their focus in the next couple years to construction of entry-level housing with fewer amenities. Demographically speaking, this would follow the need to meet the oncoming new-household creation that the data outlined above seems to suggest. And focusing on this price point might allow for a sustained recovery as the tax credits expire and low interest rates are raised.
Author: James D. Howsley, Miller Nash LLP, an Oregon and Washington law firm