Stock Buy Signals Debunked

Posted on October 17, 2012 at 12:50 PM PDT by

Clearly seeing stock buy signals is nearly a superhuman feat. If you watch enough financial cable TV shows, it’s always a small band of highly intelligent (or highly connected) insiders who seem to demonstrate a sixth sense about the market’s near-term direction.

While most of us buy and hold — or buy and hope — the gurus who profess to understand stock buy signals are always selling just ahead of the slaughter, then catching the perfect market bottom and going in big. Or at least pretending they did.

The search is endless for the magical stock market “dog whistle” that us ordinary humans cannot hear. Technical analysts, for instance, love to point out how specific patterns predict stock buy signals (or sell signals). Macro experts tout their detailed calculations of economic growth and massive demographic shifts.

Meanwhile, the fundamental crowd has an even simpler mantra: Buy stocks when they are cheap. How do we know when stocks are cheap? Why, we check the price against a company’s earnings, a figure known as the price-to-earnings (P/E) ratio.

With P/E stock buy signals, as in golf, lower is better. All else being equal, stocks with P/E ratios in the single digits or low double-digits could, arguably, be sending strong stock buy signals.

Naturally, even the fundamental analysts disagree. Consider this recent breakdown of the cyclically adjusted price earnings (CAPE) ratio from The New York Times:

Traditional P/E ratios compare stock prices with one year’s worth of historical or projected earnings. CAPE, on the other hand, looks at 10 years of averaged, or “normalized,” profits. Because this method tends to smooth out earnings anomalies that arise at the peak of a profit cycle or the depths of a recession, CAPE is considered a more conservative gauge.

Sometimes, the differences between traditional and normalized P/E’s can lead investors to drastically different conclusions.

Based on the past 12 months of earnings, for example, the Standard & Poor’s 500-stock index has a trailing P/E of around 15, which would make the market attractively priced based on historical levels, according to market strategists.

By contrast, the market’s CAPE reading is nearly 22. Although that’s not as elevated as in 1929 or ’99, it is significantly higher than the market’s long-run average of around 16.

So, is the market cheap or expensive? It’s probably better to think about stock buy signals the way doctors treat test results. A number that is within a range is given a bit of “fudge factor.” You should mention it to the patient (as in, “your cholesterol is a tad high”), but let’s not wheel him or her into the ER quite yet.

P/E ratios as stock buy signals follow a similar logic. They are most useful when they are way, way out of whack. A stock with a number in the triple digits is unlikely to ever grow earnings fast enough to match the frothy price level.

Likewise, a business that you know to be sound that is reporting reliable earnings but sporting a crazy low P/E is, in many cases, likely to see its price rise to compensate. That’s the underpinning of value investing: Find a beaten-down firm with quality management, buy big, and wait for the market to catch up. When (and if) they do, that’s when you sell.

We like the idea of the CAPE ratio. It’s logical. Most economics data is in time series that adjust for cyclical effects year over year. Nobody in the field looks at 12 months of numbers for a conclusion. Years of data is what science seeks, then they work hard to find ways to prove it wrong, to make it “falsifiable.”

So here’s the quandary: You can’t falsify earnings. If a company says it made $100 million in operating profits over the last three months, well, you have to accept that number. You can trust it or dismiss it, but you have no way to prove it wrong. All earnings figures involve a leap of faith, however small.

Looking for a way to extend the numbers out decades, even centuries, is the work of a few highly respected academics, such as Yale’s Robert Shiller and Jeremy Siegel at Wharton. Both are recognized for their serious work in the science of the markets, and understanding their ideas can be a boon for any investor interested in stock buy signals.

Stock buy signals for real investors

Nevertheless, using long-term data means that you have traded in a dog whistle for a chorus or worse, a symphony, of dog whistles. The moving parts of the market — the trustworthiness of corporate earnings, the rationality of anonymous buyers and sellers — easily distort reality. Looking for stock buy signals in the short- or medium-term data is high risk. Looking for long-term evidence is academically interesting but, well, you missed the trade.

We return to the heavy type on the bottom of every prospectus: “Past performance is no guarantee of future results.” No kidding.

What’s the solution here? Build an uncorrelated portfolio of varied asset classes and let the market fight it out. When emotions run too high in one or another asset, you get to sell and take the cash. Likewise, when irrationality reigns, you have that cash and incoming dividends to buy the value plays.

Passive investing is not really that passive. It can be pretty exciting to buy a great company or an interesting hard asset when the market has bizarrely sold off, or to cash out of an appreciated position knowing that you won a round. You can drastically lower risk, too, by doing so with cheap, highly diversified index funds or ETFs. A well-designed portfolio with periodic rebalancing is what makes it possible. No PhD required.