Chuck Ponzi March 19th, 2007

There is an old saying that if a lie is told enough times, it becomes truth. For the lack of a better term, there is a definite misconception I have seen floating about the internet that has begged for a response from someone - if not in the mainstream media, at least a blogger to attack it. Since noone has bothered to rid themselves of the problem, I guess I’ll do my best at continuing to dispell myths. (Calculated Risk already did some here)

I enjoy a good thriller as much as anyone else. Oftentimes, the truth is stranger than fiction. This myth, however seems to be created by CNN.com and perpetuated (or at least commented on) by one of the most interesting bloggers that I know of. Interesting not because he brings any specific knowledge to the table, but rather that his biting remarks, poor temper, and lack of in-depth research leaves him to be the butt of numerous internet jokes. Yes, we are talking about Larry Nussbaum. His internet presence has been immortalized by other bloggers declaring “You’ve been Nussbaumed” whenever his presence is made known by his constant truthiness comments repeated over and over enough times at least someone must actually believe them now. His most recent article that I have seen floating around is not actually a new one. It originated a number of months earlier, but the basic jist of the article is that there appears to be some uncanny similarity between the NAHB Index and the S&P 500 Index. The following even appears a number of times in different places on the internet accompanied with the following story:

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. 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.

It turns out that the mood of builders is a terrific stock market bellwether: The correlation between current builder confidence and future stock market returns over the past ten years is downright 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.

This is concerning from a number of aspects. If it were true, this would represent a significant predictor of future stock market direction.

However, the BigNose.com takes on the trouble and determines that there is perhaps only a cursory and temporary relationship.

If only there were a Sesame Street song “Correlation Is Not Causation”. I bet I’d sing it all the time.

“One of These Things Is Not Like the Others” will just have to do.

Drawing them up myself (now with even more recent data, here is how it looks (previous decade and updated to include current times). There is often too much of a bias to simply reflect our own beliefs (be it bullish or bearish) into the information we read. Minds seeking the truth will be more interested in facts that have been very different from the “expected outcome”

The First one is recreating their baseline to ensure that we have a clear picture of what is going on:

Yep, correlation exists with the base data.  Then, I tried once again the previous 10 years to see if there was a correlation:

None that I can see.

Then, I extended it out to the most recent data:

Sorry, not seeing it anymore.  We have clearly diverged in a statistically significant manner, and I’m just not clear what would cause that if there were some correlation.

Kinda reminds me of this picture at least in part.

I guess you can see what you want to see. To me it just looks like a piece of toast.

The worst part of the entire comparison is that you’re comparing apples and oranges. The NAHB is an absolute measure between 0 and 100. The S&P500 shares no such tether, and frankly with US monetary inflation the way it is, there is really no upper limit. Correlation is definitely NOT causation in this case, nor does it last very long.  Like the normal cheese sandwich that will crumble and get moldy, the CNN article just doesn’t stand the test of time (or history for that matter)

• Mean Reversion , Psychology