Current conditions in the housing market reflect a considerable imbalance between supply and demand. Much has been said about the supply part of the imbalance (i.e., overbuilding), but relatively little has been said about the demand side part of the problem. This article examines household formations and their behavior in response to the Great Recession. The main objective is to estimate the role of declining demand in what is typically characterized as the current excess supply of housing units. The primary conclusion is that two thirds of at least one recent estimate of the number of excess vacant housing units in the existing housing stock can be attributed to a steep decline in demand during the Great Recession.
Household formations (e.g., adult children leaving parents’ households, singles leaving shared housing arrangements, etc.) are the largest component of demand for additions to the housing stock. These new households are accommodated by additions to the housing stock when vacancy rates are low, and are absorbed into the existing vacant stock when vacancy rates are high. Since 1965, the number of households in the US has grown at an average annual rate of 1.5%, adding an average of roughly 1.3 million new households per year, according to the Census Bureau’s Housing Vacancy Survey.
There is a strong trend component to growth in the number of households, but formations are influenced by economic conditions, rising during good times and declining when the economy is weak. Declines in household formations can be seen in 1982, 1983 and 1990, all recession years, and in 2003, following steep stock market declines. The decline in both the economy and household formations during the Great Recession (December 2007 - June 2009) was particularly steep. Formations averaged 421 thousand annually from 2007 through 2009, roughly a third of the long term average. In this analysis a baseline is constructed using trend growth in households and pent-up demand is estimated as the deviation from that trend in the observed formations.
There is some technical difficulty measuring household formations due to the way the data is collected. Households are defined as occupied housing units (vacant units are not households) and formations are defined as the change in the number of households. The most accurate accounting of population and households comes from the decennial censuses. In the years following decennial censuses, estimates of population and household growth are produced by the Census Bureau’s Population Estimates Program and added to the prior census totals. The accuracy of the series of intercensal estimates is revealed with the following decennial census.
In 1999 the series of intercensal estimates yielded an estimate of 104.9 million households. The decennial census produced an estimate of 102.6 million households in 2000. The Census Bureau recognizes that the intercensal estimates overstated growth during the decade, rather than the alternative that the number of households declined by 2.3 million between 1999 and 2000. This difference was extreme relative to other years, but illustrates that survey procedures and revisions introduce discontinuities between some years and so caution must be used when considering shorter time frames. In Figure 1 estimates for formations are omitted where adjacent years are non-comparable.
To avoid discontinuities and capture the trend most relevant for recent history attention is restricted to the period (and data) beginning in 2000. Calculating the trend in household growth between 2000 and 2007 yields an average annual growth rate of 1.0%, slightly above the 0.8% average rate in the 1990’s, and below the 1.4% average rate in the 1980’s and the 2.3% average rate in the 1970’s. This pattern is consistent with the deceleration in growth that followed the peak in the early 1970’s as the baby boomers came of age, and a re-acceleration that followed the trough in the 1990’s as the echo boomers entered adulthood. Figure 1 illustrates the approximate consistency of the growth rates over ten year periods and the noticeable decline after 2007. Extending the 2007 level of households by the 1.0% average growth rate for the period provides a baseline against which to measure the decline in household formations during the Great Recession. Subtracting the Census Bureau’s 2010 estimate of households from the baseline yields an estimate of 2.1 million household formations that have been postponed for economic reasons.
This exercise suggests that a considerable portion of the excess housing supply (NY Fed president William Dudley recently estimated 3 million units) is due to a steep decline in demand related to economic conditions, rather than due purely to overbuilding. This insight has important implications for recovery in the housing market. If it was the case that the excess housing supply was “pure” supply (i.e., assuming no pent-up demand), then recovery in the housing market would take as long as it takes for demographic forces to “catch up” with supply (think birth rates 20-30 years ago). But, as this analysis indicates, the excess supply reflects or embodies significant pent-up demand, implying that recovery in the housing market will come more quickly as the economic recovery makes progress and pent-up demand turns into realized demand, absorbing vacant units in the existing stock and adding pressure for the construction of new units.
For more information about this item, please contact Robert Denk at 800-368-5242 x8698 or via email at firstname.lastname@example.org.