By Mark Hulbert
Is the P/E ratio worth following?
That would be an important question at any time but has become especially so now that the price-to-earnings ratio is declining.
The drop, caused by a combination of improving earnings per share and a falling stock market, brings the P/E ratio to within shouting distance of suggesting that equities are fairly valued.
That’s a big contrast to the message of the other valuation indicators I follow each month in this space (see table below), each of which continues to paint a picture of an extremely overvalued stock market.
Unfortunately, stock market bulls are on shaky ground drawing too much comfort from the P/E ratio’s recent decline. On average since the 1980s, the indicator has been no better than random in predicting the stock market’s subsequent return.
The only reason I continue to include it in the table below is because its track record prior to the 1980s was impressive —so good, in fact, that the P/E ratio’s track record over the past 150 years remains statistically significant even after including the last 40 years of randomness.
What happened in the mid-1980s to sabotage this previously good indicator? In search of an answer, I dug deeper to see if the P/E ratio continues to work at the level of individual stocks. The answer appears to be “no.”
As you can see from the accompanying chart, below, the P/E ratio’s ability to differentiate good and bad stocks declined markedly in the decade of the 1980s and then again after 2005. The P/E ratio’s track record has been particularly poor in recent years, with the 10% of stocks with the highest P/E ratios beating the lowest-ratio decile by more than 13 annualized percentage points over the past five years.
In other words, for several decades now the P/E ratio as a stock-selection tool has been more wrong than right: The stocks it thinks should do the worst have in fact performed the best, and vice versa.
This lengthy period of bad performance makes it difficult to explain it away. If the poor performance had occurred only in recent years, we might be able to attribute it to the Federal Reserve’s policy of keeping interest rates artificially low. High P/E companies’ valuations are disproportionately dependent on profits they may earn many years into the future, and lower interest rates mean the present value of those earnings will be higher. Low P/E stocks, in contrast, are relatively less affected by changes in interest rates.
But how, then, are we to explain why the P/E ratio performed as well as it did in the decade of the 1990s and in the early aughts? Interest rates also declined over that period, with the 10-year Treasury yield falling from over 8% to around 4%, and yet low-P/E stocks on balance still outperformed high-P/E stocks.
To be sure, as you can also see from the chart, the P/E ratio’s track record has suffered through long periods of underperformance before — even if none has been as extreme as the past five years. So it’s possible that the indicator at some point in the future will once again enjoy a long period of successfully identifying better and worse stocks. In the meantime, however, it remains largely a mystery as to why it’s done as poorly as it has.
Current status of valuation indicators
The table below reports the latest values of the eight indicators I report each month in this space. These are the eight that, according to my research, have the best long-term records predicting the stock market’s 10-year returns. As mentioned above, only the P/E ratio is even close to indicating the stock market might be fairly valued.
|Latest||Month ago||Beginning of year||Percentile since 2000 (100 most bearish)||Percentile since 1970 (100 most bearish)||Percentile since 1950 (100 most bearish)|
|Buffett ratio (Market cap/GDP )||1.97||2.10||2.10||99%||100%||100%|
|Average household equity allocation||50.5%||50.5%||50.5%||99%||98%||99%|
Mark Hulbert is a regular contributor to MarketWatch. His Hulbert Ratings tracks investment newsletters that pay a flat fee to be audited. He can be reached at email@example.com .