Select Page

Every month new data comes out regarding Existing Home Sales and New Housing Starts. Housing sales and starts are expected to play a large role in the timing of any economic recovery.

As these numbers come out month after month however its easy to get confused because they don’t provide historical data other than the previous months and year’s stats.

Below is my best attempt at summarizing the current housing situation in the United States. Per my usual, its a lot of charts and graphs and circles and arrows.


Total U.S. Housing Stock 2009
Year U.S. Population Total U.S. Housing Stock Owner Occupied Vacant Homes: (Sale/Foreclose/Vacation)
2009 307 million 130.4 million 74.9 million 18.7 million
2008 304 million 129.9 million 75.7 million 18.6 million
2007 301 million 128.2 million 75.6 million 17.4 million
2006 298 million 125.8 million 75 million 16.4 million
2005 295 million 124.3 million 75 million 15.9 million
2004 292 million 74 million
2003 290 million 121 million 72 million 15.2 million
2002 288 million
2001 285 million 119 million 72 million 12.8 million
2000 281 million 120 million 71 million 14.3 million
1999 115 million 69 million 15.5 million
1997 million 112 million 65 million 12.5 million

(Source) and here.

The 2050 Population Projection for the United States is 419 million. At the 2005 rate of 2.27 people per housing unit, the United States would need approximately 184 million housing units in 2050.

That figure alone should put to rest the idea that the United States has any sort of long term housing glut. If no houses are destroyed (unlikely) then the United States will need to build approximately 1.3 million houses per year for the next 41 years.

American housing stock is fairly new. Approximately 34 percent of the units have been built since 1980, and the median construction date is 1970. This median means that half of the housing units are less than 35 years old. Houses are also durable: 9.3 million units (7.5 percent of all housing units) date from 1919 or earlier and are more than 85 years old.

2005 Housing Stock Composition

Using 2005 as a sample year, the United States housing stock is typically in use, as follows:

Type Units Percentage
All 124,377,000 100.0
Year round 120,544,000 96.9
Total occupied 108,901,000 87.6
Owner occupied 74,950,000 60.3
Renter occupied 33,951,000 27.3
Vacant 11,643,000 9.4
Seasonal 3,834,000 3.1


About 60% is owner occupied, 25-30% is renter occupied, and the remaining 10-15% is either vacant, for sale or used only seasonally.

New and Existing Home Sales
Year New Households Annual Housing Starts Existing Home Sales
2009 458,000 (est.) 4.6 million (est.)
2008 1.1 million 892,000 4.9 million
2007 0.7 million 1.4 million 5.7 million
2006 1.3 million 1.8 million 6.5 million
2005 1.6 million 2.2 million 7.1 million
2004 1.0 million 2.0 million 6.8 million
2003 0.6 million 1.8 million 6.2 million
2002 2.9 million 1.7 million 5.6 million
2001 1.7 million 1.6 million 5.3 million
2000 0.8 million 1.6 million 5.1 million
1999 1.3 million 1.6 million 5.2 million
1998 1.5 million 1.6 million 5.0 million
1997 1.4 million 1.5 million 4.4 million


A few general observations:

  • There were more housing starts in 1972 (2.4 million) than in 2005, the peak year of the recent boom. Five of the ten years of that decade had housing starts of at least 2 million. New household creation averaged 1.7 million from 1971-1980.
  • 2009’s projected total of 460,000 housing starts is less than half of the next lowest housing start total in the last 42 years. Housing starts have been above one million for every year during that time.
  • Using a rough average of the 2000-2002 numbers, the United States should have been expected to average approximately 1.6 million new households, 1.6 million housing starts and 5.3 million existing home sales annually throughout the rest of the decade.
  • The dearth of new home building over the last two years has meant the total number of new homes built over the last seven years (10.6 million) is 600,000 fewer units than would have been predicted (11.2 million).
  • During the same time frame however the number of existing home sales is still way out of wack. 47.4 million existing homes were sold between 2003 and 2009 (est.). That’s 10.3 million more than should be expected.
  • To the extent that there is/was a temporary housing glut, it was created slowly over the course of decades.

The market absorbtion rate on new rental units “rented within three months” fell from approximately 72% in the late 1990’s to 62% by 2004. It was 51% in 2008.

The rental vacancy rate was ~5.5% in the early 1980’s and ~7.5% throughout the 1990s. It jumped from 8.4% in 2001 to 10.2% in 2004. From 2004 to present the rental vacancy rate has hovered between ~9.7 – 10.1%.

The Housing Glut

Clearly, even as the United States’ homeownership rate was going up in the late 1990s and early 2000s, the rental vacancy rate was an indicator that perhaps too many houses were being built. Too many houses were being bought by speculators.

A study conducted by real-estate researcher REALTrends Inc., in concert with Harris Interactive, found that real estate investors account for 22% to 28% of all home sales (existing and new) each year – a total of 1.5 million to 1.64 million houses each year.

The bubble obviously broke, but how far are we from a return to normal?

Year All Delinquencies Homes in Foreclosure Homes For Sale*
2009 9.1 million 3.85 million 3.8 million
2008 6.4 million 2.75 million 4.5 million
2007 5.1 million 1.4 million 4.4 million
2006 4.4 million 1.0 million 3.7 million
2005 4.3 million 1.0 million 2.7 million

Estimates vary, but the current shadow inventory of foreclosed homes that are being held off of the market by banks is estimated to be around 600,000 additional homes.

We’re expecting 4.6 million in annual existing home sales this year. Between 1998 and 2002 the annual average was ~5.2 million. Even after the economy starts to revive, there is clearly a large number of houses on the market (in many markets). Is there anyone out there left that wants to buy a house?

Homeownership Rates by Age

Between 1982 and 1997, overall homeownership rates in the United States ranged between 63.9% and 65.7%. In 1998, for the first time, home ownership climbed above 66%. Between 1997 and 2004 the rate climbed from 65.7% to 69.0%.

Where did these new homeowners come from? Not from the ranks of people 35 and older.

The homeownership rate for peoples 35 to 64 has remained reletively constant throughout the last two plus decades. It fell among all groups from the early 1980’s through the early 1990’s but rebounded in the decade to 2005.

Age Group Ownership Rate 1982 Ownership Rate 2005
30 to 34 Years 57.1 56.8
35 to 44 Years 70.0 69.3
45 to 54 Years 77.4 76.6
55 to 64 Years 80.0 81.2
65 and Older 74.4 80.6


Among those age groups in society that are most likely to be settled down and own a home, only two have higher homeownership rates in 2005 than the corresponding group did in 1982.

A few points:

  • The “natural” homeownership rate, if given a lifetime to build your castle, would appear to be in the high 70’s to low 80’s.
  • Some small part of the overall rise in homeownership rate would appear to have been the result of the fading away of the generation which lived through the Great Depression as adults. What I mean by that is, the 2005 “65 and over-Greatest Generation” cohort is probably more economically similar to their children (2005 45-65) than they are to their parents (1982 65 and over cohort).
  • Not listed above but relevant, the homeownership rates for whites has gone from 70% in 1993 to 75% in 2008. Homeownership rates peaked among blacks at 49.7% in 2004 and among Hispanics at the same number in 2007. HUD doesn’t provide the numbers, but it would very interesting to see a breakdown of minority homeownership by age group. As the United States becomes more “brown” either minority homeownerhips is going to have to rise, or the overall rate will fall.
Homeownership Rates for Gen-Y

So, if the overall homeownership rate went up and the 30 and over cohort wasn’t responsible, then there must have been some massive increase in ownership among those under thirty.

Was there?

Year Less than 25 25 to 29
1982 19.3 38.6
1992 14.9 33.6
2002 22.9 38.8
2005 25.7 40.9
2009 23.9 37.2


Yes. No. Maybe? Clearly homeownership rates went up among Americans under 25 in the early parts of this decade. But by 2009, those rates are only slightly above where they were in 1982.

Incredibly, young American’s 25-29 have fallen behind their counterparts from 1982.

If anything, these numbers should put to rest the idea that the housing bubble accellerated the age at which young people are owning homes. Importantly, this means that there are still a lot of renters out there who will be out there looking to purchase a home when the time is right.