Monday, May 7, 2012

If it is all about age, what should the homeownership rate be?

Last week's report of a continuing drop in the homeownership rate to its 1997 rate of 65.4 percent made me decide to revisit a question I looked at in a paper some years ago: holding demographics constant, would should the homeownership rate be?

I don;t think anyone would argue that 1990 was a year in which the homeownership rate, at 64.2 percent, was unnaturally high.  In fact, the rate had been stuck around 64 percent for around 20 years.

One of the reasons for this is that the country at the time was moving through a period with lots of young adults.  In 1990, the homeownership rates by age were as follows:

15 to 24 years 17.1%
25 to 34 years 45.3%
35 to 44 years 66.2%
45 to 54 years 75.3%
55 to 64 years 79.7%
65 to 74 years 78.8%
75 and more years 70.4%

Source: 1990 Census of Population and Housing

At the time time, the age shares of household heads at the time were:


15 to 24 years 5.5%
25 to 34 years 21.6%
35 to 44 years 22.2%
45 to 54 years 15.6%
55 to 64 years 13.5%
65 to 74 years 12.5%
75 and more years 9.2%


Source: 1990 Census of Population and Housing

Now let us look at age shares in 2010.  They are:


15 to 24 years 4.6%
25 to 34 years 15.4%
35 to 44 years 18.2%
45 to 54 years 21.3%
55 to 64 years 18.3%
65 to 74 years 11.6%
75 and more years 10.6%


Source: 2010 Census of Population and Housing

If we apply those population shares to the homeownership rate by age in 1990, we get an "age-predicted" homeownership rate of 67.0 percent.  This suggests that the rate has fallen below where it "should be."  Why might this be?  Let's look at the ownership rate among non-hispanic whites and everyone else in 1990:


White (non-hispanic) 69.1%
Non-white or hispanic 44.6%


Source: 1990 Census of Population and Housing and my tabulations.

In 1990, 80 percent of household heads were non-hispanic whites; according to the 2009 American Housing Survey, 70 percent of household heads that year were non-hispanic white.  If we hold ownership rates by race constant (as well as age) and allow race/ethnicity of household heads to vary, we would have seen a decline in the ownership rate to 61.9 percent.  Age by itself therefore pushed up the rate by 2.8 percentage points, and race/ethnicity reduced it by 2.3 percentage points, so if the effects of age and race stayed constant (and we ignore interactions of age and race for now), we would expect the ownership rate in 2010 to be about 64.7 percent.

I do wish to emphasize that the difference in ownership rates across races and ethnicities is NOT acceptable; one cannot explain difference by just looking at such things as economic status and marital status.  My reading of the literature is that African-Americans and hispanics continue to suffer from discrimination in the housing market.  All that said, it is not difficult to explain why the ownership rate is returning back to where it was.  I also think there is no reason to believe it will settle at a much lower rate than where it is right now.