Professor Robert Shiller at Yale says that the US stock market may be in nosebleed territory. “It looks to me like a peak,” he tells Yahoo Finance. Using his cyclically adjusted price-to-earnings ratio (CAPE), he notes that the current reading of roughly 26 has been higher only three times since the late-18th century: 1929, 2000, and 2007.

Shiller reminds that those three years of even higher CAPE levels reflected high points for stocks that preceded sharp corrections. The warning is timely and certainly merits serious consideration if you’re sitting on tidy gains from equities. But deciding if we’re at another moment when stocks are flirting with a new crash requires more effort than simply looking at CAPE’s elevated levels of late and deciding that the end is near.

Although CAPE and comparable metrics (dividend yield, for instance) are valuable tools for projecting expected return, it’s worth looking at the numbers in context by considering the investment alternatives. As a simple example, let’s review some of the choices during the last two market peaks that Shiller cites–2000 and 2007–and compare valuations with two other asset classes at those points in time: 10-year Treasuries and real estate investment trusts (REITs).

In the first case, December 1999 was the high point for CAPE, which reached 44.20. That was an astonishingly high valuation level, and even by today’s standards it looks like a classic case of a bubble. Nothing before or since has brought the market close to that elevated terrain. In retrospect, it was easy to assume the worst, although CAPE was only one of several metrics warning that the stock market looked pricey in the extreme.

Consider that the benchmark 10-year Treasury Note was yielding over 6% in December 1999. The dividend yield on stocks, meanwhile, had slipped to unprecedented lows at that point: around 1.2%, according to Shiller’s data. It didn’t take a rocket scientist to conclude that bonds were an attractive alternative to equities at the time. Indeed, by today’s standards, a guaranteed 6% return in a Treasury Note vs. a 1% dividend yield for stocks looks like a gift from god that screams: rebalance in favor of fixed income.

The current yield for equity REITs looked even more enticing at 1999’s close. Using NAREIT data, REITs were trading at a relatively rich trailing yield of 8.70% in December of that year.

Moving on to 2007, the CAPE peak arrived in May of that year, topping out at nearly 28. Although CAPE was elevated relative to recent history at the time, it was still well below the December 1999 peak and so by that measure it wasn’t clear in real time if stocks were overvalued to the point that a crash was imminent. Further muddying the waters was the fact the market’s 1.7% dividend yield at the time, although low by historical standards, was a bit higher vs. what was offered in December 1999.

Treasuries in 2007, by contrast, still offered a premium over stocks, albeit a smaller one compared with 1999. The 10-year Note yielded 4.75% in May 2007, or 300 basis points over the dividend yield for equities. REITs looked attractive in relative terms as well, yielding 3.8%, although the premium over equities was moderate vs. the data at 1999’s close.

In other words, the signals weren’t as stark in 2007 for deciding if rebalancing out of stocks was timely, at least not in May 2007. Stronger warnings would arrive in due course, however. In the early months of 2008, a mix of macro and market indicators were flashing red, well ahead of the carnage that arrived in the autumn of that fateful year. In March 2008, I wrote that a new recession looked like fate, a view that was less-than-universally embraced at the time.

Reviewing CAPE’s value as a leading indicator in 2007, its worth is only obvious in hindsight. Having dropped sharply from the previous peak, one could reasonably argue (as many did) that a CAPE in the mid-20s was a prudent reflection of equity valuations given the-then bullish outlook. We now know that the optimism was wrong, but everything’s obvious in the rear-view mirror.

As for the current CAPE reading of 26 in 2014, how does the metric compare? It’s certainly above average in historical terms. So too is the recent return on US stocks. The trailing annualized five-year total return on the S&P 500 is currently around 19%, or roughly twice its long-term performance. These numbers suggest that it’s timely to take some profits. The question, of course, is what do you do with the proceeds?

Cash is still yielding zero and the 10-year Treasury yields around 2.6% at the moment; REITs are a bit higher with a yield in the mid-3% range. But with recession risk low at the moment and the possibility that economic growth is picking up, the case for dumping stocks entirely and moving to an aggressive risk-off allocation isn’t particularly compelling. Yes, this can and will change, and so you should remain wary of letting today’s analysis influence your view tomorrow. The markets and the economy are dynamic systems, which implies that we should reassess our outlook on a regular basis.

As for today, what’s a strategic-minded investor to do? First, don’t get caught up in the game of reading headlines and letting the latest media interview plant fear (or greed) in your head. Dramatic changes to portfolios in one fell swoop are almost never a good idea. Why? Because our confidence level for predicting returns is generally quite low.

Even CAPE, for all its value as a forward-looking metric, comes with plenty of baggage. For an eye-opening review of why CAPE’s record for predicting returns isn’t as encouraging as it appears in 60-second sound bites, see William Bernstein’s new book–Rational Expectations: Asset Allocation for Investing Adults, pages 91-96. The problem, he explains (with the help of some academic research) is that “we’re dealing with a nonstationary system.” In other words, out-of-sample predictions rarely compare favorably with the in-sample data. Nothing surprising about that. Indeed, the caveat applies to virtually all valuation metrics, by the way, and so CAPE’s limitations aren’t unique.

Nonetheless, you’d have to be crazy to ignore the fact that Mr. Market’s dispensed hefty returns over the past several years. Returns are mean reverting in the long run. The implication: When your profits are above-average, that tells you something about risk management choices in real time. Does it mean that stocks are going to crash today? Maybe, maybe not. But even posing the question in this way is misguided.

Deciding when stocks, or any asset class, are ripe for relatively high (or low/negative) expected returns is more art than science. The risk of being wrong at any single point in time for any one asset class is high. Fortunately, we don’t have to limit our investing decisions to, say, next Tuesday at 1:37 pm and sticking exclusively with US equities. A far better plan is to diversify across the major asset classes, keeping an eye on how prices evolve, which drops clues on the how and when to periodically rebalance the mix. Even here, the results can and often do vary widely, depending on the details of the assets and the rebalancing rules.

As for CAPE, Shiller’s warning is timely. After five years of a bull market, the historical record looms ever larger for managing expectations. But the shifts in the asset weights of your portfolio should have already been dropping clues about the appropriate course of action.

Ultimately, most if not all of what you need to know about portfolio management for your portfolio arises from the shifting weights in your asset allocation. As topics for media interviews go, that’s a rather boring discussion. No wonder, then, that bubble talk and crash conversations are so popular.