What Does Dr. Doom Say About U.S. Home Prices?

If you missed the Saturday 28 Feb New York Times, then you missed their somewhat tardy recognition of the Irish economist, Morgan Kelly. The Irish call him “Dr. Doom.”

A couple of years ago, when I was developing home price models, I stumbled upon Professor Kelly’s work. I was attempting to construct a home price forecasting model that was a little more plausible than the then-typical “trend in drag.”

Professors Edward Glaeser and Joseph Gyourko (of Harvard and Penn, respectively) had summarized the then-conventional US wisdom regarding home price corrections in their “Housing Dynamics” working paper. They relied upon 1980 – 2005 data.

Looking at regional home prices using FHFA/OFHEO data, they had found that once real (i.e., inflation-adjusted) US home prices reach their peak:

• Real home prices gave back 32% of the increase that occurred in the five years prior to the peak; and
• This real decline takes five years.

While fine, as far as it went, the problem was that data that ended with 2005 could not, in my opinion, have much to say about post-2005 corrections, since the pre-2005 increases were much smaller than the post-2001 boom, which had not yet corrected.

We were living, so to speak, “way out of sample.”

To see this graphically, take a look at the following two charts.

Figure 1: Real FHFA/OFHEO and S&P Case Shiller National Home Prices

Figure 2: Annual Change in Real FHFA/OFHEO and S&PCS National Home Prices

They depict, respectively, the level and annual changes in each series, where the green line marks 2005, where Professor Glaeser & Gyourko’s FHFA/OFHEO data set ended. While there had been corrections prior to 2005, these corrections were give-backs of increases that were much more modest than the post-2001 boom.

While I was comfortable with the Glaeser-Gyourko metric (i.e., give back of a portion of the increase that occurred 5 years prior to a peak), I was unhappy with the US sample that they had to use.

Where could I find research that worked with increases similar to that of post-2001 America?

And that’s when I discovered Professor Kelly’s On the Likely Extent of Falls in Irish House Prices. Using an international (as opposed to US) data set, Professor Kelly wrote:

• Looking at house price cycles across the OECD since 1970, we find a strong relationship between the size of the initial rise in price and its subsequent fall…

• Economic theory … predicts that house prices should not follow a random walk, but should be a mean-reverting process of booms and crashes around a slowly increasing trend reflecting the growth of household income. This is what the international data show…

• Typically, real house prices give up 70 per cent of what they gained in a boom during the bust that follows. This is a remarkably robust relationship, holding across very different OECD housing markets over more than 30 years.

Needless to say, this dire forecast was a tough sell a couple of years ago – the correction was more than twice as severe as the then-conventional wisdom.

While I managed to use re-expressions of this forecast in models that I constructed of total mortgage market originations (which seemed pessimistic at the time but turned out to be accurate), Dr. Doom’s grim anticipation of a “70% correction” seemed outlandish.

Now, of course, his then apocalyptic view seems quite rosy. Given the flatline trace of household income, we are headed back to what will be a 100% give back. That will bring us back to the real home prices that prevailed in 2001. About fifteen percent lower than today.

Comments

[Tom Armistead:]

The Federal Reserve put two scenarios on their website to be used for the stress test on the bank. Their baseline projection is for a little more than 15% from today, the stress case is for 25% from today.

As you say, Dr. Doom's past projections are now a rosy point of view. I like the 15%.

Mar 01 06:56 PM

[The Last Bo...:]

I am not a housing expert. Just a regular guy who tries to decide if the home prices are low enough in order to buy a house. Is it safe to get back to the market as a buyer?
I used data from the OFHEO monthly seasonally adjusted index of home prices in the US. The data is from 1991 through 2008. The index rose with an average of 0.2-0.3 from the beginning of 1991 through the beginning of 1998. In April 1998 the index is 123 (100 in January 1991). Then it accelerates and rises much faster until 2006. In 2006 the growth slows down and after reaching a high of 224 in April 2007, the index starts decreasing and it is 199 in December 2008.
The interesting question is: where the index would have been if its growth rate were the rate it had in the period 1991-1998? I recalculated the index using this historical growth rate and its theoretical value in December 2008 was 155. This is much lower than the actual level of the index: 199. In the scenario that excludes the housing boom the index level of 199 should have been reached in 2023. But unfortunately it is still only 2009 and we have index of 199 suggesting that the housing prices are still way too high. My conclusion: I am not buying. It is too early to call the bottom.

Mar 02 12:17 AM

[Ira Artman:]

Dear Last Boomer:

Thanks for reading and note. Your analysis, which constructs a trend by avoiding the bubble, is not bad. If I had to quibble, it would be with your choice of starting year (forced upon you by series you selected, I believe), which is 1991.

Think that was the beginning of recession (from my living memory), so think your trend - beginning as it were,in a recession, could be low due to your choice of start.

But idea of "bubble avoidance" is good one/metric to develop. - Ira

Mar 02 07:12 AM

[John Lounsbury:]

Ira - - -

Don't you think the trend line would be artificially steep if you start from an unusually low point (recession)?

Mar 02 10:48 AM

[Ira Artman:]

John Lounsbury

Thanks for reading and note. Not in this case.

Look at Figures 1 and/or 2.

Last Boomer is using a series that begins in 1991 and ends in 1998. He picks up post 91 weakness and then ends series in 98, well before or just as the market began to revive. So for practical purposes, most of his 91-98 period represents what seemed like a housing bust, for those of us that lived through it, prior to current.

Make sense? - Ira

Mar 02 11:15 AM

[John Lounsbury:]

Ira - - -

When I was a child I used to look at the pictures in the National Geographic and skip the text. I just got caught doing the opposite.

Mar 02 08:10 PM

[Hmm?!:]

Ira,

What about the incredibly lax lending standards recently ....I do not see a way that could be figured into your trending. But I think it is a major factor this time around.

(Mortgage Liquidity Du Jour by Credit Suisse,
March 2007)
"low/no documentation loans increased from just 18% of purchase originations in 2001 to 49% in 2006"

"sampling 100 stated income loans found that 60% of borrowers had
"exaggerated" their income by MORE than 50%"

That means 30% of ALL mortgages in 2006, the buyer only had 66% of
the income they claimed. Other mortgages were also likely bad from
minute one. If the buyer lied by 49% or 30% on their application?

I just cannot see a way this meltdown burns off, for quite some time.

Mar 03 12:58 AM

[Ira Artman:]

re Hmm?! , about lax lending, etc.

Think that would account, to large degree, for the explosion and collapse in the S&P Case Shiller index - lending of sort you describe was largely for non-conforming (non Fannie/Freddie/Ginnie) loans. Correction I focussed on in piece was for FHFA/OFHEO.

Thanks for comments. - Ira

Mar 03 11:04 PM