Unraveling the Housing Market
By: Dr. Charles Lieberman
Date: 3/18/2009
Estimates of the time needed for the housing industry to work off its excess inventory range from sometime within calendar year 2009 to outside estimates of four to five years, a rather wide range for such a crucial sector that is key to the economic and investment outlook. After a lengthy and careful review of the data, as well as the analysis offered by analysts of widely different perspectives, it is our conclusion that the housing market is likely to hit bottom in 2009. Therefore, housing should contribute to an economic turnaround in the second half of this year.
Overview The collapse in housing that began in 2006 has reduced GDP by roughly one percentage point at an annual rate for nearly three years. Housing price declines have also depressed household net worth, which has undermined consumer spending. The decline in housing values has also contributed to large losses to investors when households default on their mortgages. The resulting loss of capital at financial institutions has curtailed their ability to lend or invest in the economy. So, the severe weakness in housing has inflicted considerable damage on the economy through a wide variety of mechanisms. And at least a modest recovery in housing is necessary to relieve some of the stress in the financial system.
An examination of the excess supply of housing hanging over the market suggests that considerable progress has already been made and that the housing market is approaching bottom. There are four markets where the inventory excess is still sizeable, notably around Las Vegas, the Inland Empire region of California, Phoenix and parts of Florida and the housing news from these markets should remain fairly negative for some period of time. Elsewhere, the excess inventory is considerably lower. On a nationwide basis, we expect new home construction to flatten in the months immediately ahead, exerting little additional drag on GDP. As excess housing inventories decline, home values should begin to rebound, a necessary condition for home builders to be able to resume construction and earn a profit. The decline in housing inventories will be the key data that will predict the turnaround in housing.
Estimating Excess Housing Supply There is a very wide range of estimates of the excess supply of housing hanging over the market, from a few hundred thousand units to a few million units. In an editorial in yesterday's Wall Street Journal, LeFrak and Shilling suggest that the excess supply of housing is 2.4 million units, which if correct, would take several years to absorb, if new construction remains severely depressed. In fact, the excess inventory of the nation's housing stock is poorly captured by the most commonly cited inventory figures of housing on the market i.e. new and existing homes for sale. It is our thesis that the overhang of excess housing inventory is captured by the sum of above normal level of unsold new homes plus the above normal level of vacant existing housing. We will work through the sources of excess housing, as well as sources of demand, to estimate the glut that has depressed housing activity. Table 1 provided below summarizes our approach and can also serve as a framework to the analysis that follows. (All of the figures should be considered rough estimates.)
Table 1
Excess new housing inventory 0
Excess vacant existing housing 750,000
Excess rental housing 350,000
Overestimated housing stock -1,500,000
Second home supply 750,000
Net inventory overhang (rough) 350,000
The first component, newly constructed homes for sale, is the unsold level of new construction. To the extent newly built home inventories are above normal, that excess is clearly part of the glut. With new construction activity severely depressed, these inventories have been falling for almost two years. Over more than twenty years, unsold new housing inventory has ranged between 300,000 and 400,000. With the inventory down to 342,000 as of January 2009, it appears this supply is now at a level that might be considered normal. So we ballpark the excess portion of this inventory at around zero. (Compositionally, there is plenty of excess around the four overbuilt markets, but that is offset by below normal levels in other areas.)
The harder part of the calculation is to discern the abnormal quantity of existing homes for sale, which is the larger part of the puzzle. The total number of existing homes for sale is not a good measure of the excess supply. When a seller sells, they must move elsewhere, either into a home purchase or a rental. Such relocations neither add nor detract from the inventory of units on a net basis. They merely reallocate the existing housing supply. However, some existing homes for sale are already vacant, as occurs when a seller has already purchased another home and moved. In this case, the unsold, vacant unit is part of the inventory overhanging the market and it is just as much of a burden on the market as an unsold newly built home. Thus, the above normal level of vacant existing homes for sale is part of the glut that must be worked off for the housing market to return to normal.
Since the Census Bureau publishes quarterly data on home vacancies, it would seem a simple matter to examine this data to assess the degree to which vacancies are above normal. A review of the historical data suggests that vacancies are about 1% too high, roughly 750,000 units. Vacancies of rental units also appear slightly above normal, but this stock is much smaller, so the excess inventory is likely around 350,000 units. Putting together vacancies from all sources, new homes, existing homes, and rentals, the entire overhang appears to be around 1 million units.
In fact, the quality of the Census data on vacancies is quite poor. The estimates of the total housing stock for 2007 and 2008, is based on trends from 2005 and 2006 (as noted in the footnotes to the press release www.census.gov/hhes/www/housing/hvs/hvs.html), although these were peak years for housing and home construction. So the Census Bureau estimated an increase in the total housing stock of more than 2.2 million units in 2008 from 2007, using an above normal growth rate based on 2005 to 2006, while actual housing starts in 2008 were just over 800,000. By relying on the boom years to project out the nation's housing stock, the Census Bureau overestimated the total by a considerable margin. A comparison of the estimate of the growth in the housing stock to actual housing starts suggests that the housing supply is over-estimated by at least 1.5 million units, based just on the lower level of actual new housing construction.
The weakening economic activity has also surely added to the supply of vacancies from atypical sources. The weak economy is forcing some home owners to unload second homes or vacation homes because they are under financial stress. There are 4.8 million vacant seasonal homes according to the vacancy data. But in these difficult times, some of those homes have surely been put up for sale. If 10% (a guess) have become available for sale, an additional 450,000 units of vacancy would be added to the inventory overhang. But it is possible the actual figure could be twice as high. So, we added 750,000 units to the housing glut estimate from the supply of second homes placed on the market.
Aggregating all of the factors listed above suggests that the true overhang of excess housing inventory is about 350,000 units, but the discussion associated with each component of this estimate should provide a good sense that the estimates are subject to considerable uncertainty. It is possible the actual figure could be twice as high as the 350,000 figure, but it could also be lower.
Housing Absorption Any assessment of the housing glut must also account for trends in new household formation, which is needed to absorb this inventory. Underlying demographic trends are very favorable for a rapid rate of household formation, since there is a very large population cohort exiting college, entering the job market, and getting married. These are the children of the baby boomers, in effect the echo of the baby boom, and they are a very large cohort. Our population is growing by about 3 million annually and U.S. household formation should be between 1.3 and 1.5 million annually. In fact, this trend estimate of household formation is surely too high for the moment, reflecting the financial constraints on young people that prevent them from moving out on their own in a weak job market. Instead, college grads may move back home or share housing units with other new graduates until economic conditions improve. Immigration has also surely slowed. Unfortunately, we don't have very good estimates of how much household formation has slowed down in response to these cyclical pressures. Therefore, the absorption rate of the (unknown) inventory excess is also subject to considerable uncertainty. I will use half the underlying demographic rate of household formation to ballpark this figure, so roughly 700,000 new households being formed in this weak economy.
Table 2 Household Formation Trend 1,400,000
Overestimated household formation 700,000
Cyclically-adjusted household formation 700,000
So how long will it take for the excess to be absorbed? Assuming household formation is about half the underlying rate suggested by demographic trends, so roughly 700,000 households annually, would imply that the excess inventory would be absorbed within about one half year. If the actual glut is higher, or the rate of new household formation is lower, it may take as much as a full year to absorb the excess housing inventory.
Some allowance must also be made for the concentration of vacant housing by geography. If vacant housing is concentrated in Las Vegas, Phoenix, Florida, and the Inland Empire region of Southern California, other areas will start to experience a tightening housing market before those overbuilt markets return to equilibrium. Thus, new construction might need to increase in Texas, the Carolinas, and Northern California before the entire excess is worked off. Watching the behavior of new construction in these markets may provide some insight as to when the housing sector will stop being a drag on overall economic growth. Similarly, location also matters for the incremental supply of vacation homes, since this housing is unlikely to be in the same geographic areas as primary homes and, thus, exert less of a supply deterrent to new construction in cities. The same point can be made even with regard to new primary home construction in the distant suburbs of Los Angeles, Las Vegas and Phoenix, where vacancies may take quite a while to fully absorb without exerting a major influence on the price or sales of homes in the closer suburbs. Thus, the inventory of unsold new homes may exert less of a drag on housing than would be suggested by an inventory level of 342,000 units.
Reality Check How do these estimates jibe with the actual behavior of the housing market? It does appear to be very likely, even compelling, that significant absorption of the excess inventory is occurring, since both the supply of new homes and existing homes on the market has been falling. The inventory of new unsold housing has fallen from a peak close to 600,000 units two years ago to about 342,000 units as of January 2009. Unsold new housing inventory is now near the mid-point of its long-term historical range, yet enjoys better demographic trends. This is an unambiguous substantial improvement. Existing home inventories have plummeted by more than 20% over the past six months, providing some reinforcement to the implications of the decline in new housing inventory. Qualitatively, this decline is a positive development, although it is impossible to distinguish between occupied and vacant units, so its true importance is difficult to assess.
Conclusions Our calculations suggest that the housing supply imbalance is falling at a good pace, which leads to the conclusion that the housing sector is likely to bottom this year. The uncertainties surrounding key aspects of the analysis precludes us from reaching a firm conclusion that the bottom will be reached within six months, although such an outcome appears reasonably likely. Even with the wide range of uncertainty surrounding these estimates, it appears quite unlikely that the housing sector will require an additional 3 to 4 year time frame to complete its adjustment, as some are forecasting, barring a dramatic further downturn in the economy.
Monitoring the progress being made in eliminating the excess supply of housing is best done by tracking housing inventories, sales and new starts. Inventories of unsold new homes should decline somewhat further, although the pace of decline should slow now, as inventories in the four badly overbuilt regions sell off slowly, while other regions begin to recover. A continued rapid pace of decline in the inventory of existing homes for sale during the key spring selling season would strongly suggest a nearer term end to the housing glut. However, a rise in the supply of homes for sale might be very difficult to interpret, since more homes may come on the market once owners locked into their current homes believe the market has improved enough for them to be able to sell. Any pick-up in housing selling rates would be a distinct positive. A flattening or pick-up in new home starts, as was reported for February, would also be a distinct positive, since it would signal that builders are running out of inventory and still see enough demand to resume some construction activity. Download this article in PDF Format
About Advisors Capital Management, LLC
Advisors Capital Management, LLC (ACM) is a provider of privately managed portfolios and financial planning services for industry professionals and direct clients. Although the information included in this report has been obtained from sources ACM believes to be reliable, we do not guarantee its accuracy. All opinions and estimates included in this report constitute the judgment as of the dates indicated and are subject to change without notice. This report is for informational purposes only and is not intended as an offer or solicitation with respect to the purchase or sale of any security. ACM is a registered investment advisory firm. For program fees and descriptions please request a copy of the firm’s ADV part II Schedule F. Web Address: www.advisorscenter.com 777 Terrace Ave, Hasbrouck Heights, NJ 07604 Phone: 201-426-0081
©2005 Advisors Capital Management, LLC 
