It is generally agreed that housing is a significant contributor to the current economic crisis. Many have said that the credit bubble can not be deflated and the consumer can not return to spending until housing prices stabilize. Others have said that recovery will not be sustainable until the housing market bottoms. I have heard so many people offer opinions (watch CNBC) that the housing market will bottom in 2009 or 2010. A few have offered an opinion of a longer wait. I have not seen a comprehensive discussion of the factors involved in the housing shake-out, so here goes. I hope it’s not a case of “fools rush in where wise men fear to tread.”
There are three aspects of the housing crisis: Falling prices, mortgage defaults and supply issues. They will be tackled one at a time, although they are related. Falling prices increase mortgage defaults, which throw foreclosed houses onto the market, and thus the inventory glut is increased. Additionally, there is a section analyzing the demographics involved in the current housing cycle, which are different from the previous cycles since 1950.
There is a famous graph shown below, widely reproduced in the media and around the internet. It even made it to 60 Minutes TV. What the graph shows that there is a bigger ARM (adjustable rate mortgage) problem ahead than the sub-prime problem, which is in its final stages. The potentially most problematic part of the coming bubble (from the default point of view) is the Option Adjustable Rate mortgage. These have optional payment provisions, in some cases allowing zero payments, until the reset date. All shortfalls from the normal amortization payments are added to the principle. It is like compounding interest in reverse. The mortgagee is compounding the principle upward.
My eyeball estimate from the graph is that these mortgages probably comprise around 20% or so of all the ARM resets in 2010 and 2011. If these mortgages were issued at 80% to 100% of 2006 market prices, and the current value reflects the current national average price decline (around 25%), plus some of the interest payments have been deferred, how likely is default when the reset is made at a higher principle than originally contracted? This is very likely, especially if the low payment option has been used due to limited ability to pay.
According to Credit Suisse, there are 8.1 million foreclosures expected in the four years 2009 – 2012. The projection was made before the dramatic unemployment increases reported in December. If unemployment increases more than assumed in the analysis, the estimate for foreclosures could increase.
This estimate of 8.1 million foreclosures is approximately 42% of the estimated 19 million homeowners that have been estimated will be underwater (owe more than the house is worth) by next year (David Leonhart in The New York Times, The Trouble With a Homeowner Bailout). This 19 million represents about 50% of all mortgages. In the same article an estimate is quoted from Mark Zandi, Chief Economist at Moody’s Economy.com that during the life of their mortgage, 6.5 million will not be able to afford their mortgage payments. The Credit Suisse estimate passes the sanity test if these two other numbers are accepted. If most of the 6.5 million not affording their mortgage payment default by 2013, and if most of them are in the 19 million underwater by 2010, then we only need to have 11-14% of the remaining underwater mortgagees default to get over 8 million.
The Alt-A and Option Adjustable Rate mortgages resets are relatively low in number in 2009, but rise dramatically in 2010 and even higher in 2011. If resets could be made with current interest rates, some of the resets could be managed by mortgagees that would default at higher rates. However, it could be argued that the only way for the current low rates to continue for three or four years would be for continued deflation, much further fall in house prices and much higher unemployment. These factors could overwhelm what might otherwise be advantages from lower interest rates.
The biggest factor in mortgage default will probably be affordability of payments, but walking away from a big mortgage with negative equity may also contribute significantly. The same Credit Suisse report estimates that sixty-two percent of Alt-A borrowers (up from 41 percent in September) and 83 percent of option ARM borrowers (up from 66 percent in September) are projected to have negative equity in the next two years. We will discuss the prospect for house prices over the next few years in a later section.
If the average mortgage balance in the 8.1 million defaults is $300K, the total value of all mortgages defaulted would be more than $2.4 trillion. Assuming a 50% recovery after all foreclosure, maintenance, insurance and property tax expenses, the loss to lenders is $1.2 trillion; with 60% recovery, $1 trillion. If the average mortgage in default is $240K, scale the preceding numbers down by 20%.
Finally, this mortgage foreclosure overhang will add approximately 20 months of inventory at current home sales rates. Current new housing starts are approximately equal to the new home sales rate. If we assume that this balance is maintained, then the current inventory of unsold homes over 11 months (Greg Robb, Economic Report: U.S. data show sales of new homes at lowest level in 17 years), plus the estimated foreclosures represent over 2.5 years of inventory. This is an extreme headwind for the building industry to overcome. The prospects for new housing starts and sales are discussed in a later section (House Supply Outlook).
House Price Outlook
There is a saying: “This time it’s different.” This is often criticized as being wrong, that current events have similarities in the past. Well this time is different. House prices have gotten away from historical relationships with construction costs and affordability. The effect of these imbalances is seen in the graph below comparing decline in house prices in the current cycle compared to the last cycle in the 1990s.
This graph shows a comparison of the percentage change of the Case-Shiller Home Price Index for the housing correction beginning in 2005 (red) and the 1989–1990s correction (blue), comparing monthly CSI values to the peak value seen just prior to the first declining month all the way through the downturn and the full recovery of home prices.
The graph below shows how out of whack home prices have been recently when compared to construction costs. The graph stops before the house price index top in the summer of 2006 above 226.
Source: Robert Shiller, “Irrational Exuberance”, Princeton University Press, 2005
The following graph shows how house prices and family incomes have varied over the past 25 years. The black reference line gives the reference indicating where the average house price has to come to get back to the historical relationship with income. The data is taken from the Case-Shiller Housing Index and the U.S. Census Bureau data base. The 2008 estimated family income is not available as of this date, so the 2007 data point was repeated for 2008 estimated average family income. To reach the historical reference, average house prices would need to fall to the $120,000 to $130,000 area over the next few years. This assumes incomes continue to grow as they have over the past 25 years. This price target is $40,000 to $50,000 below the December average price (23 -28% lower).
On the graph a target box is drawn showing a price bottom target between early 2010 and late 2011. Of course, if the price decline is steeper than the dotted projection, the bottom target would move in toward the end of 2009. If the drop is more gradual, the projection would move out into 2012 or 2013.
Another estimate of how much more house prices have to drop can be found from data in an article by Ira Artman (What's in Store for Regional Banks? ). Abstracting data from Ira’s chart entitled “US and Metro NYC Prices, Actual & Forecast”, we see a projection of sharply slowing house price losses over the next 18 months. My projection from that chart is house prices 10-15% lower, with a bottom by the end of 2010 in the range of $150,000 to $160,000.
Jim Kingsdale (Deflation Watch) has projected the bottom for houses won’t arrive until 2013 and at average prices around $110,000 (2007 dollars). If we have any net inflation between now and 2013, Jim’s price estimate is not that much different from my nominal price range.
Looking at price to rent ratios, Matt Blackman (Housing Prices: Bottom or Temporary Bear Break?) estimated in September that house prices had about 25% further to fall at that time. Adjusting for price declines since, and assuming the same rental cost now as then, that leaves about 18% further price decline by this metric. (Note: In the text of the article, Matt suggests a 12% drop in house prices is indicated to return to the historic ratio, but my reading of his graph yields 25%.) Applying a linear extrapolation to Matt’s graph would project a bottom in the second half of 2009.
The following quote is from an IMF working paper worth reading in its entirety:
In our best judgment, single-family home prices as measured by the OFHEO purchase-only index were around 14 percent above equilibrium in the first quarter of 2008, with a plausible range of 8 to 20 percent.
We find that the bloated inventory-to-sales ratio, high foreclosure rates, and the large degree of inertia in housing markets imply that recent price declines are likely to continue. Moreover, with the gap between actual and equilibrium home prices playing only a weak anchoring role, the downward momentum could well take home prices considerably below equilibrium.
Source: IMF Working Paper WP/08/187, “What Goes Up Must Come Down? Housing Price Dynamics in the United States” by Vladimir Klyuev (pdf warning)
Although the IMF estimate of over-price levels in early 2008 now appear to be too small, the assessment that there is no guarantee that over-correction will not occur is worth noting.
There may be other estimates of what price and what date will define the bottom of the housing market, but I will move on to the supply issue.
House Supply Outlook
It was pointed out (in the mortgage section earlier) that the current house backlog is over 11 months and that the coming wave of mortgage reset induced (and other) defaults could add another 20 months. This creates a potential sales headwind of 2 ½ years of sales inventory before another house is built. If new construction continues to contract to match shrinking demand, the additions to the sales inventory may come primarily from foreclosure in the coming years. Let’s get to the outlook.
In the graph below is shown the house sales volume (per 100 million of population) over the past 45 years. We have indicated a target for the bottom of this cycle, at about 90,000 units per 100 million of population. This is about 275,000 units at the bottom projected to occur between late 2009 and 2011.
From the graph above we can calculate the new single family home sales (annualized) estimated for the fourth quarter, 2008 is 427,000 (140,000 x 3.05 (hundred million multiplier)).
The next graph displays the total residential housing starts since 1958. It indicates the bottom projected from previous cycles has already been exceeded.
According to the U.S. Census Bureau (Quarterly Starts and Completions by Purpose and Design), in 2007 (latest data available), 65% of the total residential housing starts were single family. This implies that in the fourth quarter 2008, about 410,000 single family units started construction (annualized). This number comes from 220,000 (from above graph) x 3.05 (100 million multiplier) x .65 (fraction which are single family). This is very close to the estimated single family residence annual sales rate, 427,000, calculated previously.
If the bottom of the housing market sees sales of only 275,000, sometime in late 2009, 2010 or early 2011, as estimated previously, there must be additional decline in housing starts by about 33% (135,000 units annualized) sometime during the next 1-2 years.
One additional note: Steven Hansen (http://seekingalpha.com/article/104904-it-might-not-be-possible-to-stop-the-decline-in-housing-prices) has reported that the actual number of house that could be in the current inventory, but are not, is at least as large as the 11 month inventory reported (and probably larger), because many “discretionary” sellers are waiting for a better market.
The biggest demographic factor on the table is the baby boom. The front wave of the baby boom is now in their early sixties and the wave continues for 18 years. The baby boomers were born from 1946 to 1964, so in 2009 they will be the age group from 45 to 63. In the graph below (from here) vertical lines have been drawn at five year intervals to define the 20 years containing the surviving baby boomers.
As the baby boomers move to retirement, downsizing and increasing mortality, there is a smaller group of successors to occupy the houses of the elders. Increasing mortality has been depleting the front edge of the baby boom but the number of houses they occupy has not been diminished proportionately because of surviving spouses.
The nature of the baby boom housing overhead can be estimated by looking at the younger edge of the boom, now 45 to 50. It shows populations 350,000 to 450,000 greater each year than the fifteen years following (30 to 45). If we add back 75% of the decline from age 50 up (my estimate of the survivor effect, which may be too low), the estimated baby boom households for age 50 to 55 becomes close to the 45 to 50 total. For 55 to 60 the household total is estimated to be 100,000 to 150,000 lower than the approximately 2.5 million to 2.8 million estimated households for the ages 45 to 55 (4.5 million divided by 2 plus an adder for single resident occupied housing). The range in estimated households results from uncertainty in the number of single baby boomer residents – the total of spousal survivors, the separated and the never married.
So my estimate is that the housing market for the fifteen years following the baby boom is 150,000 to 200,000 households less per year than the youngest 15 years of the baby boom. That is a 15 period with demand loss per year which is 35% to 50% below the fourth quarter 2008 sales rate. This 15 year period will be coming in the downsizing peak period of the baby boomers, when some will be moving to smaller houses and wanting to shed second homes.
Steven Hansen (A Tale of Two Housing Bubbles) has written a recent article that reports a rapid growth of second homes (vacation homes and “investment” homes) over the past 10 years from1-2 million in 1997 to about 20 million as of 2006. (This is only includes sales from 1997 forward, so the 20 million may be understated. On the other hand this number does not appear to be corrected for resales, so the 20 million may be overstated.) This is about 15% of all homes in the U.S. These are much more likely to be added to the housing market (offered for sale or foreclosed) than a primary residence when the owner has financial distress. Also, the increasing burden of dealing with two or more homes as one gets older will be an incentive for baby boomers to get back down to one house just at a time when demand for houses is falling for demographic reasons. At current single family home sales rates, 20 million is more than a 46 year inventory!
One final demographic note comes from Spengler (The Devil and Bernard Madoff ) who recently published the graph below.
Prospective home buyers and home sellers in the US:
Spengler argues that U.S. demographics will create a crossover to more people over 50 than 20-50 by 2015. He misleadingly calls the younger group buyers and the older group sellers, while there may be buyers and sellers in both groups. However, he does argue, with merit, that this does not bode well for the future value of large homes, which are less preferred by the older group.
There are some conclusions worth presenting.
- It has been estimated that there will be 8.1 additional foreclosures by the end of 2012. This is 21% of all mortgages.
- It has been estimated that 19 million mortgagees will owe more than their house is worth by 2010. This is 50% of all mortgages.
- It has been estimated that about 6.5 million mortgagees will not be able to afford their mortgage payments at some time during the term of the mortgage. This is 17% of all mortgages.
- It is possible that homes for sale inventories, currently over eleven months, could reach as high as 2 ½ years during the next four years.
- The average additional drop in average home prices (five sources) is estimated to be 20% to reach the bottom.
- It has been projected that new home starts will fall to approximately 33% (from 4th quarter, 2008) to 275,000 (annualized) by the time we reach the bottom for the housing market. If this does not occur, the bottom may be lower in price and take longer to reach.
- Estimates for the time the bottom will be reached (six sources) range from late 2009 to early 2013. Four sources estimate 2010 or 2011.
- The estimates made above (1-7) might become worse (lower prices, fewer housing starts and later bottom) if negative demographic factors were appropriately included.
- The estimates made in this summary could also become worse if there is a prolonged recession (or depression) of two years or more. Conversely, if this recession were to end in the next 5-6 months, the estimates might improve.
It is the author’s overall conclusion that Federal intervention to solve the housing crisis would consume too much resource that would be better used to stimulate future economic growth. The data reviewed in this article indicates that the home building industry may never return to the levels of the past decade unless demographics are changed in an unanticipated way by increased immigration or increased birth rates.
Disclosure: The author has a large home which he will probably want to sell in 10-12 years. There are no positions at present in banks, mortgage companies or any construction businesses.