Shiller P/E (does it matter?)

Coffina: One metric that I've been looking at a lot recently is the Shiller price-to-earnings ratio, and this metric uses 10-year average of real GAAP earnings in the denominator. The advantage is that it's been much more predictive historically of future total returns than one-year price-to-earnings ratios, which really hold little to no predictive value. A one-year price-to-earnings ratio will tell you, for example, that the valuation looked in line with historical norms right up to the financial crisis, and then it actually peaked in late 2009, after the recovery was already well under way.

In contrast, the Shiller price-to-earnings ratio was at relatively high levels, about the 80th percentile relative to the last 25 years, or in other words, it had been lower 80% of the time in the lead-up to the financial crisis. And then in the depths of the financial crisis in late 2008 and early 2009, the Shiller price-to-earnings ratio was seeing levels that it hadn't seen in 20-plus years.

So, the Shiller P/E tends to be a lot more predictive of future total returns, and that measure right now is at about 26, which is about the 68th percentile relative to the last 25 years. So, in other words, the market has in general been cheaper, based on this measure, 68% of the time since the late 1980s. That's certainly a cause for concern based on recent history.

So, what we've seen historically is that any time the Shiller P/E is above, say, 25.5, which is about the 60th percentile relative to the past 25 years, subsequent total returns for investors have been very poor, on average in the low-single digits, with some very severe drawdowns

Glaser: So should investors then be preparing for a big sell-off, or a repeat of 2008?

Coffina: It's very, very hard to say. It's one thing to say that the market is richly valued relative to historical standards. It's a completely different thing to say that this means stocks are going to decline over the next one year, three years, or even five years. As I mentioned, in 1996 and in 2002, the S&P 500 was valued at similar levels and went on to have a great run over the next five years until it ended in a crash. So even over a time period as long as five years, saying that the market is relatively richly valued relative to history doesn't tell you much about what stocks are actually going to do in the near term.

Glaser: Over the next 10 years, then, from these valuation levels, what kind of total return should investors be thinking about?

Coffina: Well, that's a great question, and I wish I had a better answer for you. It's going to largely depend on what happens to price-to-earnings ratios. I mentioned before that over the last 25 years, the Shiller P/E ratio took a big step up versus the prior 100 years. The measure used to be around 14-15. Over the last 25 years the norm has been more like 23-24. So anyone who is looking at a measure like this in, say, the early '90s would have decided that, say, 1992 was a great time to sell stocks, and they would have missed out on the '90s bull market, and even through the 2002 crash, they would have missed out on total returns in the high-single digits.

There's definitely a caveat to this whole kind of analysis; market conditions can change, and past is not necessarily a good indicator of what's going to happen in the future. In particular, in the environment that we're in right now, if interest rates stay as low as they are currently, that can certainly justify sustained higher valuation levels going forward. I'm not saying that's going to happen, but it's certainly possible.

[8/24/14 Mitch Zacks chimes in]

Currently the CAPE ratio is flashing some warning signs as it hits levels that have previously been seen only before some of the major market sell-offs of the past century. Interestingly enough, however, valuation multiples based on trailing twelve-month earnings and future earnings estimates are showing a market which is expensive but not unusually extended given current interest rate levels.

The question for investors is, of course, whether the high reading of the CAPE ratio should prompt a reduction in equity exposure.

Now, I am a little biased as Zacks created the quarterly consensus earnings estimate which effectively enables valuation multiples to be calculated based on forward looking earnings data. P/E multiples based on analysts' earnings estimates show a market that is more expensive than historical, but not at the eye-popping levels shown by the CAPE ratio. Despite this bias, at the end of the day, I do not think long-term equity allocation should be adjusted based on the current high reading of the CAPE ratio.

First and foremost, I have learned over many years it never makes sense to time the market. It does not work. It does not work if you react to newsletters, earnings trends, IPO activity levels, discounted cash flow models, P/E multiples, gurus, magazine articles, tea leaves or sunspots.

In fact, one of the few metrics that does seem to have some value in terms of predicting future market performance is interestingly enough, tracking what Wall Street investment strategists—the analysts who set equity and fixed-income exposure of brokerage firm model portfolios—are recommending to investors and promptly do the opposite. The reason this methodology may work is that by the time multiple strategists are calling for the market to go in a certain direction, the information they're responding to is already reflected in stock prices. As a result, the market surprises by reacting to new information and moves in the opposite direction. This rule of thumb would currently indicate increasing equity exposure as many of the strategists are recommending a reduction in equity exposure.

Additionally, there are a few problems with the CAPE ratio. All the methodologies of calculating P/E multiples are designed to compare current prices to future expected earnings. There is agreement amongst investors that the key metric in stock market valuation is future earnings.

No one buys a stock such as Apple ( AAPL ) based on what the company earned over the past twelve months. Investors instead focus primarily on what AAPL will earn in the future and whether those earnings are greater than expectations currently built into the stock's price. As a shareholder of AAPL, you are valuing the company based on the future theoretical dividends AAPL can potentially pay as opposed to the historical dividends it has paid. The past earnings data, however, is often used as benchmark for predicting what the future will bring. Want to know what Apple will earn next year? Look at what they earned last year and then estimate how many more iPhones they will sell.

The CAPE ratio makes the assumption that historical ten-year inflation adjusted earnings are a good predictor for corporate earnings over the next ten years. As a result, the CAPE multiple is showing an extremely high level because current stock prices are being compared to trailing ten-year historical earnings.

Essentially, ten-year historical corporate earnings numbers may not be a good predictor of the future as they include the years 2008 through 2010 when corporate earnings were incredibly depressed due to the financial crisis. Therefore, an investor's faith in the CAPE ratio as a predictive tool of future market performance comes down to whether or not corporate earnings will revert to historically depressed levels.

This depends to some extent on whether the financial crisis is seen as a recurring event or a once in a seventy-five year outlier. I am certain the financial crisis is an outlier as opposed to a recurring event, and as a result, the CAPE ratio is likely not as meaningful as it has been at other times.

Corporate Earnings, Inflation and Interest Rates

Unfortunately, another argument why corporate earnings will mean-revert is because corporate profit margins, which are at all-time highs, will move back to historical lower levels. This argument has a stronger leg to stand on because if wage inflation starts to pick up, corporate profit margins should come under some pressure.

As I have written numerous times before, the market's P/E multiple is higher than its historical average primarily because interest rates are substantially lower than what they have been. If interest rates remain below historical levels then the market's valuation is reasonable. It all comes down to whether wage inflation materializes. If it does, the market will be in for some rough sledding with decreasing corporate profit margins while the P/E multiples come under pressure due to rising interest rates.

Despite my anticipation of inflation as a result of quantitative easing, no inflation has actually materialized for more than two years. The trillions of dollars sloshing around the fixed income market buying and selling U.S. government bonds is telling us that interest rates are expected to remain low for quite some time. The yield on the 10-year would be below 3% only if inflation was expected to remain incredibly low for a very long time. 

Though I am convinced that once the yield on ten-year U.S. government bonds starts to revert to its historical levels, stock market valuation multiples, regardless of how they are calculated, will come under pressure. Ultimately, the future direction of the market is going to be dependent on whether interest rates remain low or inflation returns to the system.

So Where Is the Market Headed Now?

While I do not think the market is at a dangerous level of valuation, I do feel that future returns in the market will be lower than they have been over the past five years. Rates are going to have to go higher and valuations are going to have to come down.

[That's safe to say.  The past five years have returned 16.58% mainly because it doesn't include 2008 which returned -37.00%.  The returns have been (starting in 2009): 26%, 15%, 2%, 16%, 32% (for an average of 18%, 16.58% for the five years from today).  AAPL returned 147%, 53%, 26%, 33%, 8% (for an average of 53%!, 34% for the five years from today).  AAPL dropped 57% in 2008.]