Housing and Recessions: A Proposed Relationship

Floyd Norris, the noted financial columnist for the New York Times, has recently begun blogging in addition to writing his regular column.  He recently posted some research he had conducted on the housing market and recessions.

Norris takes a three-month moving average of housing starts to smooth out the volatile series.  Next he notes that we are currently in the eleventh month of decline in this moving average.  Finally, he looks back at all of the other times there were eleven months of decline, going back as far as he had data (1959).  Here is what Norris found:

1. November 1973 was the 11th month. A recession began that very month.
2. April 1980 was the 11th month.  A recession began in January of that year.
3. November 1981 was the 11th month. A recession began in July of that year.
4. February 1991 was the 11th month. A recession began the previous July.

These days, almost no one thinks a recession is looming.

While he does not quite say it, the implication is that an eleven-month decline is an indicator of a recession around somewhere.

This would be a good starter problem for an undergraduate class.  Students who had the benefit of reading (my favorite professor) Neil Browne’s book, Asking the Right Questions:  A Guide to Critical Thinking, (featured on our reading list) would find this an easy problem.  Perhaps Scott Rothbort, who recently offered some great insights about the housing market, will test this question on his class at Seton Hall.

The astute readers of "A Dash" have, no doubt, immediately spotted the main issue.  The popular bearish argument is that a decline in housing will weaken the economy enough to cause a recession.  In Norris’s examples the recessions all occurred either before or (once) contemporaneously with the housing decline.

Are we surprised that during a recession, spending on housing — and many other things– is reduced?  The housing decline is associated with the recession.  Association does not show causation.  If one had to guess, the housing decline is probably a result, not a cause in Norris’s examples, both from logic and the timing.

In short, Norris’s little exercise has the look of sophisticated analysis — smoothing of data, moving averages, the same eleven-month time period, and the fatal recession.  Although he has a big audience, I was hopeful that Mildred, one of my investors who is easily influenced by this sort of argument, would not read about this "relationship".

My hopes were dashed when I saw that Norris had already gotten a boo-yah from Barry Ritholtz on The Big Picture.   I know that Mildred watches Barry on TV and also reads The Big Picture.  I can expect a phone call soon.

No one knows when the next recession will occur, nor whether it will be induced by problems in the housing market.  We can say with confidence, however, that the Floyd Norris four-case analysis does not help us in making this prediction.

You may also like


  • RB February 21, 2007  

    On the one hand, we have the wizards on Wall Street proclaiming a bottoming of housing, perhaps from their years of experience in dealing with highly liquid assets or because they did not step away from their keyboards looking for homes to buy. On the other hand, there are the same builder CEOs who are much more candid away from earnings calls.
    Perhaps the expert with the ouija board is right after all — after all demand doesn’t care about mundane things such as pricing or affordability. I’ll instead conclude that CEOs aren’t that optimistic after all and that Goldman Sachs’ Jan Hatzius is the biggest doom-and-gloomer of them all.

  • Nova Law February 22, 2007  

    Barry Ritholtz’s usual modus operandi is to look for any news which reinforces his bearish orientation. You should not be surprised when he trumpets any tidbit which can be used to describe the glass as “half empty.”

  • PM February 23, 2007  

    “While he does not quite say it, the implication is that an eleven-month decline is an indicator of a recession around somewhere.”
    While you’re giving lessons on critical thinking, that’s a straw-man argument. Floyd Norris didn’t claim causality. The start date of recessions is only known after the fact. So it’s legitimate to look for indicators that a recession may be underway or imminent — and this is not a bad one. If your real purpose is to deny the possibility of a recession, you have lots of other indicators and statistics you can point to to show that this time it’s different.