Mon 6 Nov 2006
CNNMoney vs. Blog — Blog wins
Posted by Steve under Misc
Last week I read an interesting article that I almost mentioned here. The article, titled Can the economy survive the housing bust? was posted on CNNMoney and written by Jon Birger, senior writer at Fortune magazine. First, let’s look at how the article opens:
Tucked away in the briefcase of Liz Ann Sonders, chief investment strategist at Charles Schwab & Co., is a chart so scary she’s hesitant to show it to investors.
Holy crap! That sounds pretty scary, don’t it? I mean, surely the chief investment strategist at Charles Schwab has been around the block once or twice and knows bad things when she sees it. So what is this terrifying chart?
It plots the National Association of Home Builders’ Housing Market index - a monthly measure of builder confidence - against the Standard & Poor’s 500 stock market index, with a one-year lag.
Okay, I see. So the NAHB index is supposed to tell us something about the future direction of the S&P 500. What does it show?
It turns out that the mood of builders is a terrific stock market bellwether
Yeah, okay, I already said that. Get to the point.
The correlation between current builder confidence and future stock market returns over the past ten years is downright unnerving.
Oh? How unnerving?
Not only did the NAHB index presage the start of the post-1994 bull market in stocks, but its decline starting in 1999 foreshadowed the equity market collapse that came the following year. Builder confidence rebounded in November 2001 - a year ahead of the stock market upswing that began in October 2002.
Why is Sonders worried now? Just look at the chart.
What chart? Oh, the one below this.
Over the past year, the NAHB housing index plummeted 54 percent. Were stocks to follow suit, the S&P - 1400 in late October - would be trading below 700 this time next year.
HOLY CRAP! SELL! SELL! SELL!
I have to admit that is a pretty scary chart. To be fair, the article went on to note:
Sonders isn’t predicting anything so apocalyptic, but she doesn’t hide her concern about housing and her pessimism about stocks.
Okay, I happen to agree. While I think the economy is headed for a slow down, I think even the most pessimistic investors wouldn’t expect the S&P 500 to fall back to 700 in the next twelve months.
So that was the article. I thought it was pretty interesting, even though it was indeed perhaps a bit unnerving.
Then, today I was reading a wonderful blog called Calculated Risk and saw the aforementioned article discussed right here.
What did Calculated Risk have to say about this?
This is an example of a forced correlation. By picking a certain time period, and adjusting the scales, the correlation looks very strong.
The first thing you should note is that the CNNMoney chart begins in 1994. What the article fails to mention is that the index in question was introduced in 1985. Back to Calculated Risk:
But here is a comparison over a longer period - since 1985 when the NAHB Builder confidence index was introduced. In the long run, Builder confidence moves between about 20 and 70, whereas the S&P 500 trends higher. So the comparison breaks down over a longer period.
Okay, what does that chart (from Calculated Risk) look like:

Oops! Looks like the “correlation” does indeed fall apart in the long run. In fact, even in the last ten years, the correlation isn’t actually as strong as the CNNMoney chart suggests.So, what to make of all this? Well, as I’ve mentioned before, I think it’s legitimate to be concerned about where the economy is headed over the next year. But CNNMoney is guilty of some very sloppy reporting. The chart they show is worth discussing–and it may even turn out to show the direction of the S&P 500 over the next year–but it’s equally important to show that the recent “correlation” is probably nothing more than a coincidence.
Bad CNNMoney! No soup for you!

November 6th, 2006 at 9:05 am
Well, in general I’d recommend not taking anything “they” say too seriously. “They” aren’t usually any more successful predicting the future than “we.”
November 6th, 2006 at 9:37 am
Happy to help. I enjoyed your article and hope the boost of traffic will be good!
November 6th, 2006 at 9:55 am
> but it’s equally important to show that the recent “correlation” is probably nothing more than a coincidence.
It isn’t. As the USA has deindustrialized, housing has become a much more prominent part of the economy. Especially in the very recent past.
November 6th, 2006 at 12:29 pm
Good job busting CNNMoney. This is the reason why blogs rule and why newspapers and their sites are losing subscription. Keep up the great work!
November 6th, 2006 at 1:34 pm
One interesting note is that the S&P 500 never gets above the NAHB index. This is probably because the index is built off of many stocks that follow market patterns (almost all stock indexes have the same trajectory or downfall). Its possible that the NAHB index could be below the S&P 500 but probably not for long.
November 6th, 2006 at 1:49 pm
I have always been amazed by the impact that short term forecasting and results have on stock prices and market mood. If there a company like IBM or Microsoft were to miss their quarterly goals the whole market is depressed for a day or two. But they don’t take into account the larger picture. The same is true here. What about the fact that there is a finite number of new homes that can be placed on the market before there are no new buyers; too many homes or not enough buyers will mean a drastic drop in starts. Does this mean that there’s no buyers for stocks out there?
I think any correlation between two or more indicators has to have a long history of proof before it can be used for any form of forecasting. Unfortunately a 12 year correlation that breaks prior to that is not a good track record in my opinion.
Great article by the way.
November 6th, 2006 at 5:44 pm
Excellent article thanks
November 6th, 2006 at 8:39 pm
Good catch Steve! I guess this is why new financial aggregators like Instant Bull seem to emphasize blogs over traditional media, newspapers around the US are really beginning to feel the squeeze due to growing influence of blogs
November 7th, 2006 at 6:56 am
30 year search for historically similar charts in the SP500 (Yearly Similar Chart Search).
Observer that 1 year after the similar charts (change mode to Predictability) there is a 75% correlated slight trend upwards.
http://datawink.com/blog
November 16th, 2006 at 10:25 am
How can I plot this chart on an excel spreadsheet?
November 18th, 2006 at 4:39 pm
no soup for you!
lol that’s funny