›

how fast can the equity freight train stop?

deejayohdeejayoh
edited May 2007 in Housing Bubble
I thought it was interesting to look at the OFHEO data across markets, to see just how abruptly appreciation went down from sky high rates, and then consider how Seattle compares.

from OFHEO

ofheodatadi9.jpg
Eyeballing it, looks like 60-0 performance is about 3-4 quarters. We just hit the brakes for Seattle.

Comments

  • The_TimThe_Tim
    Oooh, fun idea. Here's a graph of some similar data. I took the Case-Shiller / S&P HPI data and lined up the "cliff" (the point where YOY appreciation peaked and began to head down for good) on each of the cities at "June 2004" on the chart.

    Seattle is the thick light green line with circles, Portland is the thick light blue line with circles, and the composite-20 (which includes Seattle) is the thick orange line.

    CS-SP_HPI.png

    City - months from cliff to negative YOY
    Detroit, MI - 11
    Boston, MA - 13
    Tampa, FL - 14
    Phoenix, AZ - 16
    Washington, DC - 18
    San Francisco, CA - 19
    Cleveland, OH - 19
    Minneapolis, MN - 29
    New York, NY - 22
    San Diego, CA - 25
    Average - 18.6
    Composite-20 - 29
    Composite-10 - 30

    The following cities have not gone YOY negative yet.

    City - Jan. '07 YOY - Months since cliff
    Portland, OR - 8.75% - 10
    Atlanta, GA - 2.30% - 12
    Seattle, WA - 11.14% - 13
    Dallas, TX - 0.48% - 13
    Los Angeles, CA - 1.03% - 13
    Chicago, IL - 2.17% - 14
    Miami, FL - 4.16% - 15
    Las Vegas, NV - 0.02% - 28

    As you can see, we're well along the way to the same territory as the rest of the country, roaring economy or not. Just because we're taking a while to get there doesn't mean we won't.

    Maybe I'll make this into a post. Thanks for the idea.
  • deejayohdeejayoh
    hmmm. interesting. I am thinking you could do a regression against all the markets for points from peak down - and then have a best fit function for the decline across markets, plus a probability model.

    I was going to use CS - but was too lazy. Just happened to be perusing the OFHEO site
  • deejayohdeejayoh
    OK Tim -
    you got me curious. I went back to the CS data, and grabbed annual appreciation for all markets. I threw out those without at least 1 year >10% because they don't seem all that "bubbly". I also left out Seattle (since it's the one I want to predict) and the indexes.

    Then I scatterplotted all of them against the period number, and ran a polynomial equation for that.

    I get y = 0.00005x^3-0.016x+0.307
    R^2=0.644 (When I had in the low appreciation markets, it was worse)

    The function shows that it takes about 21 months from peak to hit zero appreciation, and that the market will stay negative for about 12 months. By the way, I show seattle in month 15, with a max of 18.5%. not sure what the difference is versus your run.

    Fun with Numbers!

    csregressioneq9.png
  • deejayohdeejayoh
    Next rev -
    I tweaked a bit more - for this version I lost all the outliers. I just used the markets that were in the high teens to low twenties. Interestingly, that left me with LA, SF, Portland, Seattle, and DC. That's it. I used seattle in this analysis, because I needed the data. It is highlighted on the chart

    Here are the results. R-squared is up to 0.865, and the function looks a lot more sensible for Seattle. I do think we have about 8-10 more months of appreciation - then 8-10 down months. Recession would probably hasten the trajectory down and lengthen the dip.

    csregressionpi4.png
    The "CS index will be lower in 18 months" bet that I made in January is looking 50/50. I could squeak out a free dinner - but it will be close
Sign In or Register to comment.