Cyclical vs. Secular Market Forces

Last week we discussed our outlook for the S&P 500 (SPX) for 2007, which included a higher target range for the index (1550-1600), as well as for both earnings per share (EPS) (something on the order of 9%-plus earnings growth) and a rising price/earnings ratio (P/E). It occurred to me that we ought to go into how the market tends to treat EPS and PE.

Over the very long term, the market's performance tends to be strongly correlated to earnings growth. By "very long term," I am referring to a secular time frame. And in this case, by "secular" I'm specifically referring to the past 45 years.

The top pane on the first chart plots the log of the SPX price against the log of forward 52-week EPS. We use the log of these series so the data at the left-hand side of the chart is visually meaningful. If we plotted the arithmetic series, everything before 1991 would look flat and meaningless. But the important point is unaffected: The correlation between the two arithmetic series (SPX price and forward 52-week EPS) is extremely strong at +0.96 (where +1 is perfect and -1 is perfectly inverse).

That settles it, right? The market trades off earnings growth. It's as simple as that, so let's leave well enough alone and just look at market fundamentals!

That's great if your time frame is 45 years (or even 25 years). But what if your time frame is somewhat shorter than that? (As most of our time frames are!)

Look at the second pane on this chart (black line). That pane is the market's P/E on forward 52-week EPS. Pretty variable, eh? It fluctuates in the 14-18 band for most of the 1960s, drops down to as low as 6.5 in 1979 and then rises as high as 26 in 1999. How do we make sense of that? (We'll get to that in a minute.)

What I've found in my research is that while the long-term correlation of the SPX and forward 52-week EPS is extremely strong, over the shorter term the correlation is much weaker and earnings growth is a less-than stellar prognosticator of market performance.

This next chart shows us the SPX (log scale again) in the top pane and the P/E on forward 52-week EPS in the second. In the third pane, however, we now plot the SPX year-over-year change against the same change in forward 52-week EPS. And while there is a positive correlation, it is nowhere near as strong as the long-term correlation of +0.96, but rather has shrunk down to +0.19.

That's a statistically significant relationship, but not an incredibly strong one.

OK, so if EPS growth isn't the prime mover of the market over shorter time frames, what is?

Well, much more significant than EPS growth is PE expansion/contraction.

The correlation between the year-over-year change in the SPX and in forward 52-week P/E is a much stronger +0.69.

So what does that mean in concrete terms? Well, it means that one way to beat the market is to understand what drives P/E expansion and contraction. And a factor that is extremely significant is inflation. Higher inflation generally leads to a lower P/E for the market. And lower inflation generally tends to do the opposite.

Why? For a variety of reasons, but most concretely because when inflation and interest rates are lower, the current value of future earnings is discounted at a lower rate. This roughly means that when inflation and interest rates are lower, $1 earned a year from now is worth more now than if inflation and interest rates are higher. (Other reasons include the decreased predictability of real economic growth and profit margins in higher-inflation environments. And there are probably oodles more reasons to boot.)

This next chart shows the first chart above and adds both the 10-year Treasury yield (TNX) and the consumer price index (CPI) year-over-year change.

As you can see, the two spikes in CPI during 1974 and 1979 (yellow highlights) led to the formation of the lows in the market's P/E. And once the CPI (along with bond yields) dropped back down to reasonable levels (say, under +6% on the CPI), the market's PE could be seen rising again.

Hmm. But what about the downtrend in the SPX P/E (black line in the second pane of the last chart) that's been ongoing since 2000? We'll talk more about that specifically in our next discussion.

For now, though, those of us with a cyclical time frame of something less than 40 years -- say, anywhere from one to two-and-a-half years -- must be aware that the factors that drive P/E expansion/contraction will have much more to do with how the market performs in our time frame than will aggregate earnings growth on the SPX.

---

Adam Oliensis is president of Dog Dreams Unlimited, a guaranteed introducing futures brokerage, and editor of the trading service The Agile Trader. Under no circumstances does the information in this column represent a recommendation to buy or sell stocks. Oliensis appreciates your feedback; click here to send him an email.