Medusa, You Look Awful! Can't We Do Something With Your Hair?
Lo, this market is a horrible sight, isn't she?
If you’ve not read my first installment regarding the S&P500 Corrections and the Coppock Curve, I suggest you backtrack and start there. Then come back to this article so you can better understand what I’m about to lay out. Of course, if you are the type of person who only reads the instructions after you’ve discovered you can’t put the new toy together, then go for it and plunder ahead!
As promised, I have some updates to the Coppock Curve and the S&P500 Corrections graphs that I’d posted a month ago. (Sorry it took awhile but since the lottery hasn’t paid me yet, I’ve gotta keep the day job, which took me to Detroit for awhile, which leaves things at home queuing up until I got back, yada yada yada).
In the comments (#10) of the first “Sweet Medusa” post, I mentioned there is a correlation present between the Housing Market Index (HMI) and the S&P500 Index. The former is produced by the quintessentially optimistic NAHB (National Association of Home Builders). One of the reasons that MauiPeter and I wanted to wait until the 15th of July is because that’s when the Barney-imbued NAHB updates their HMI figures and charts. “I love you, you love me, please don’t bad mouth NAHB…”
In this blog, I’d like to add a little more fuel to the fire to explain why I think we’re in for more bad news to come. In particular, I want to talk about the work of David Rosenberg, the Merrill Lynch economist who discovered an interesting relationship between the HMI and the S&P500 in 2002(?). After some back-testing, Rosenberg found that the S&P500 followed the HMI fairly closely with a one-year lag. That is, the HMI could be used as a leading indicator for how the S&P500 would do 12 months into the future within some close correlation. How closely, you ask? How’s an 80% correlation factor grab you? Yep, a pretty fair following, I’d say. But, as with most back-testing techniques, there are limitations to clairvoyant applications of said data projections. Like unicorns, those crystal balls are exceedingly rare and can completely disappear if you bet the house on them. Let’s examine it more closely to see what we can learn…
The NAHB began distributing their HMI in 1985, and until 1996, the correlation between it and the S&P500 held about as much real value as the recent African rumor of penis theft via black magic. (Notice I didn’t say ‘penis snatching’ as some reports have, because that might imply a completely different sort of transformation…;-). However, since then, things have changed. As reported in this 2007 article:
“Between 1996 and 2000, the HMI increased 90 percent while the S&P 500 more than doubled in value. Between 2000 and 2002 the HMI dropped 40 percent while the S&P 500 lost 46 percent from peak to trough.”
Chartmaster MauiPeter has provided me with his own graph of this trend, as shown below. Click on the picture to get a larger view, if so desired. He’s taken the HMI graph and scaled it by 20X so that it can be plotted on the same Y-coordinate axis as the S&P500.
Click here for the full image.
Other variations of this chart have been shown on the Web with one of the two curves shifted forward (or backward) by 12 months, in order to demonstrate that the two curves line up fairly well for the period of interest. For comparison, here’s one example from late 2006 (source: TradingEducation.com):
Humans are such visually-based animals; trusting that someone really computed an 80% correlation factor pales in comparison to actually showing you the overlain graph lines. However, the data in our chart has not been shifted by any number of months so that you can see for yourself what the HMI and SPX values were for any specified date (something which you can’t do with other versions of this data on the Web, which don’t have this recent data logged either). But trust me: the correlations were computed in an Excel spreadsheet using the “CORREL” function.
The correlation held tight until the summer of ’05. Then, in early July 2005, the HMI topped out at 71.6 and has nose-dived to about 18 today. For the correlation to have held, the S&P500 should have started moving down around July 2006, but it kept right on rising (then around 1270) until October 9th, 2007 (at 1565). So, the correlation was broken then, right? I don’t think so. Not yet, and not completely.
Rosenberg found the 80% correlation to exist in data taken between 1996 and 2002 (approximately). Hard as it may be to believe, the correlation has actually grown a bit tighter recently. My friend MauiPeter fiddled with his spreadsheet and found that by shifting the SPX curve backwards by 15 months (instead of 12), the correlation improved to 85%. Those additional 3 months might explain some of the delay, but it doesn’t explain all of it.
Also, if you look closely, you’ll see that the S&P500 is back exactly where it was two years ago on July 10th 2006 at 1236.20. All that SPX value over the past 2 years has evaporated; as of August 2nd, 2008, the SPX is at 1249.01. So perhaps we’ve caught up (caught down?) with the HMI’s precipitous fall. Contributing factors for the SPX’s delayed decline have been suggested to be as varied as:
the Fed’s continued insistence of devaluing the American dollar (something by which I think everyone is quite confounded);
the extraordinary bubble into which housing was led by cheap money and avarice lending & investment house practices (which didn’t really get started until 2002), a bubble which many people think has passed by us, but it hasn't;
the rise in oil & commodity prices;
the war in Iraq and it associated governmental largesse (“War is Peace; Freedom is Slavery; Ignorance is Strength,” anyone?)
Each of these, I believe, has pushed the S&P500 curve out to the right in the chart above.
Rosenberg's work is not without its detractions, though. Most notably, the period prior to July 1996 yielded very little correlation. Who’s to say that the period we’re entering now won’t completely unwind the Rosenberg correlation? Well, you might, but I won’t. Not yet anyway. If the correlation is breaking down, it won’t do so overnight or even in 6 months. For all of these reasons, I think we’re still in for a lot of pain to come, at least as far as the stock market indices go.
Here are the updated S&P 500 Corrections (normalized to Oct. 1st, 2007) and the matching Coppock Curve through July 27th, 2008.
Click here for the full image.
Click here for the full image.
Looking at the S&P 500 Corrections curve, the black line represents our current debacle (fitting color, isn’t it?), placing the index at around 1260 now. That black line seems to be following the leads of the 1973-1974 and 2001-2003 slides, tracking one for a couple dozen weeks, then switching to track the other. If this market decline is going to be as worse as either of those two slides, we could see the S&P 500 index cut to 810, a 48% chop from the high of 1558. A shockingly low number indeed! It’s hard to believe it will fall that low. However, I do think it will fall to 1000, plus/minus 50 points, by December 2009. That’s a 21% drop from today’s value.
As you can tell from the updated Coppock Curve, we still have a ways to fall to reach the 2001-2003 low. Personally, I think this recession will be worse than that 6-7 years ago (though it’s still hard to get my head around an S&P500 value of 800!). If it’s only as bad as the 2001-2003 slide, we should count ourselves extremely lucky. The Coppock Curve, perhaps, is showing us some good news in that if (and I must bold-face & italicize this), if our current slide will be only as bad as the last one, then maybe we’re more than halfway there, and the downward slide won’t hit the “810” mark.
Ok, what have I been smoking?! Time to take off the Pollyanna hat!
I think we’re going to have a Coppock Curve like that for the 1973-1974 recession, dropping to the -30 value, meaning we have much further to go. This is why I’m pretty sure we’ll see the S&P500 index fall to at least 1000, and possibly lower. The one variable I don’t have a good grip on (amongst many, I’m sure) is inflation. An open question: How does the rapidly ramping up inflation (now estimated at 5.02% for July, based on government figures though, which we now are purposefully low-balled) affect this predicted number of mine? Note: the S&P numbers in the graph above has not been corrected for inflation either. These are straight out of Yahoo! Finance historical data.
As previous bubbles have deflated, there always seemed to be a new area into which money could/would flow, thus begetting a new bubble (in time). I don’t see many places like that right now, except perhaps energy & maybe some commodities, though commodity price increases are being driven partially by interest rate cuts and its darker side, inflation.
So here’s my second group of questions: What are you doing to protect yourself, and where do you see value in this crazy market, if anywhere? Do you think commodities stand a chance at keeping to their elevated prices, in particular copper? Also your thoughts on the above graphs would be appreciated too.
And finally: Has anyone seen Perseus? Where is he when you really need him?!