By historical standards, stocks are overvalued, hovering near their second-highest price level ever. Recently, the Shiller cyclically adjusted price-to-earnings (CAPE) ratio stood at 33 for the S&P 500 Index, double its historical average of 16. That doesn't bode well for stocks if that ratio starts to move back toward its long-term norm.
However, many serious commentators consider that the market, though somewhat overvalued, certainly does not float on a bubble. "Yes, the American stock market is expensive, maybe 20% above its normal value," considers Raymond Kerzérho, director of research at PWL Capital in Montreal. "But other developed regions are not overvalued. Canada is at 24 times earnings, Europe at 22 times, Asia-Pacific at 20 times."
Kerzérho and others think that many factors justify valuations significantly higher than the historical norm. The most obvious factors are interest rates, which are currently at historical lows, and inflation. Vanguard has devised a "fair value" CAPE that takes into consideration these factors and which arrives at a much lower market valuation than what the Shiller CAPE establishes, which is based only on past inflation. Vanguard claims that, since 1990, its CAPE is much better than Shiller's at predicting future returns.
Jean-Philippe Tarte, faculty lecturer at HEC Montréal, notes that "a high P/E ratio like the one we presently have does not predict a recession. However, we are right in expecting returns that are below the historical norm of 8%. It will be substantially lower for the next five years, but I couldn't put a number on it."
Another factor is that transaction costs have plummeted. "Today, thanks to ETFs, in a single trade I can construct a highly diversified portfolio for a near-zero cost. Only 25 years ago, that was unthinkable. To have a roughly equivalent portfolio, I would have needed to buy thousands of stocks in many countries, and the cost would most likely have been extremely high."
Two other factors most certainly play a role, though they can't be quantified. The first is that "in the last 140 years, we have moved from a rural to a developed, and now to a hyper-developed economy," points out Robert D. Arnott, founder and chairman of Research Affiliates, LLC, in California. "As it evolves, you find that there wouldn't be as great a risk premium attached to investing."
One significant example is the Internet and how, linked to global trade networks, it contributes to keep inflation in check, according to Steve Rogers, investment strategist at Investors Group in Toronto. "I can buy something from a supplier in Mongolia rather than from one down the street," he says.
The other significant factor Arnott identifies is demography. There's an army of baby-boomers who have no choice but to set aside money for their old age "and they are valuation-indifferent," he says. "They can't care for what the price is for an asset because they have to set money aside. They are also 'sell' indifferent: they have to sell to get revenue, no matter what the price or yields of stocks are."
All these factors combine to create a very different playing field where valuations can't align with the historical norm. And Arnott expresses the view of many when he says: "All this suggests that the right valuation (for the U.S. market) would be in the low twenties. So we think that prices should correct by 20% to 25%."
Though not necessarily, because prices are probably already in the zone Arnott delineates, according to Rogers. That hinges on the sky-high valuations of a few hi-tech stocks – the so-called "FAANGs" and a few others. For example, 
Amazon.com (AMZN), with a trailing P/E of 260, accounts for 2.8% of the S&P 500 capitalization. "If you could eliminate just four of those top stocks, you would come close to historical averages," he claims.
All those factors still leave open a key question: what is a fair price?
Some types of investors don't really care for "fair value", notably momentum-style investors and, very often, growth investors. But for value-style investors -- a pretty large group -- it is crucial and amounts to determining the "intrinsic value" of a stock. Damien Conover, director of health care research at Morningstar in Chicago, is such an intrinsic value miner. As he explains, finding that value involves performing an analysis of future cash flows and discounting them by the average cost of capital. Of course, identifying future cash flows can involve loads of details: competitive dynamics, industry-specific costs, revenue projections, management features, etc.
The "intrinsic value" such an analysis delivers serves to estimate if the stock's price is high, low or... fair. Yet, that can still lead to almost opposite perspectives on the market. Two value-style portfolio managers to which I recently spoke perform approximately the same discounted future cash flow analysis. One concludes that he can't find low prices anymore and is hoarding cash in preparation for a coming correction; the other one still sees attractive opportunities.
Such apparent contradictions hinge on features that hold considerable uncertainty, explains Steve Rogers. "Each company has unique characteristics that make its future uncertain," he says. Fair value is hard to nail down simply because the future is not a neat picture you can hang on a wall.
Analysts and portfolio managers usually perform an intrinsic value analysis on a company-by-company basis, but boosting that analysis to encompass the whole S&P500 could be highly informative. It could supply a better perspective on market prices because, contrary to Shiller's index which is backward-looking, it would be forward-looking. And that is very nearly what Morningstar has set up with its "market fair value" ratio, which calculates a fair value ratio for all the stocks that Morningstar analysts cover. It's not quite a total market index, but it still covers thousands of stocks. Now, lo and behold, standing at 0.93, that Morningstar market fair value ratio presently confirms Steve Rogers's view that markets are really not that expensive (at 1.00, the ratio would consider that stock prices are just right).
A better way to deal with "fair value" could be to think in terms of "rational expectations", as Jean-Philippe Tarte suggests. "A fair price often appears as just any price that the market is ready to pay. But that doesn't mean the price is rational."
Such a "rational price" could present itself as a price range that would support "rational" expectations of future returns. Quite interestingly, Robert Arnott addresses the issue from the perspective of bubbles. A bubble, he claims, is in force when "you need to make implausible aggressive assumptions about future revenues to justify present valuations. Apple's (AAPL) stock price is not in a bubble, because you don't need implausible assumptions to justify its price. But for Twitter (TWTR), Netflix (NFLX) and Tesla (TSLA), you need implausible assumptions."
