Understanding How to Analyze Market Metrics

Longtime readers of mine — going back to the email days 10 years ago — will note that at various times I have been bullish or bearish (or both, depending upon different sectors). Regardless of our investment posture, there is always a good faith effort to maintain an analytical rigor and intellectual discipline in our investment approach.

This means sometimes discarding attractive theories, or embracing counter-intuitive ones. It also means questioning your assumptions, stress testing your theories, and whenever possible, having your ideas "peer-reviewed" to identify weaknesses or possible theoretical holes. Blogging has been very helpful in that regard.

That approach also means reading a variety of different viewpoints, theories, and people — most especially those who I may be in disagreement with; It forces a rethink of current principles, and keeps the reasoning process sharp.

Which brings me to a column on CNBC.com today. In it, Vince Farrell mentions towards the end:

"Jason Trennert, my oft quoted strategist, calculates that dividends on stocks
exceed the yield on 3-month Treasury bills. That has happened two other times in
the last 50 years and both coincided with major market bottoms."

That seems to be fairly reasonable on the surface . . . but after thinking about it for a few minutes, it struck me as potentially having some hair on it. 

A little digging, and the flaws in the reasoning surfaced:

Start with where Real (not nominal) Rates are today: They are currently negative at 3.0%. That makes the 3 Month T-bill rather artificially low. That becomes a second variable worth reviewing: Does this buy metric — SPX dividends > the yield on 3-month Treasury bills — work when real interest rates are negative?

The quick test of this theory is to see when the last time Fed engineering might have pushed the 3 month rate below the dividend yield.

It turns out that they had done so recently. Indeed, the dividend yield on the S&P500 has been above the 3 month T-bill for quite a while now — since February 2008. (Recall the emergency Fed cuts made just prior to that). If you made any purchases on the basis of SPX Yield exceeding 3 Month T Bill, you ended up with some jumbo losers on your hands.

Hence, we must repeat our prior admonishment against relying on Single as opposed to Multiple Variable Analysis in making any market forecasts or buy decisions. More often that not, you can find varying degrees of correlation, but end up lacking a causation sufficient to make an informed investment decision.

I would suggest those who like to use this metric to rerun their analysis — but run a cross check with a 2nd variable of real versus nominal rates. You may find a rather a different set of results . . .

S&P500 Yield (blue) Vs 3-Month Treasury Yield (red) against SPX (gray)


chart courtesy of Michael Panzner


Single vs. Multiple Variable Analysis in Market Forecasts (May 2005)

“A Whiff of Panic . . .”  (January 22, 2008)

Farrell: Touching Bottom in this Bear Market?
Vince Farrell
Jul.22 2008, 7:23 AM ET