The time to buy has always been when others are despondently selling. Valuations on stocks look very compelling right now, with the S&P 500 trading for just 11.85x 2011 and 10.75x 2012 earnings, after Friday’s plunge.
Put in terms of earnings yields, we are looking at 8.44% and 9.30%, while 10-year T-notes are only at 1.92%. The old “Fed Model” suggested that the forward earnings yield (call it 8.80%) should be in line with the 10-year note. Instead, we have the dividend yield on the S&P 500 higher than the 10-year note.
If we exclude the companies that pay no dividend (and thus yield 0% regardless of where they trade) then 83.6% of all S&P 500 stocks yield more than the five-year, 60% yield more than the 10-year, and 34% yield more than the 30-year Treasury. With the possible exception of the very depths of the 2008/09 crisis, such valuations have not occurred in my lifetime.
There is a counter argument that based on 10-year average trailing earnings, stocks are somewhat higher than the very long-term historical average. On that basis, the P/E is at 19.45x while the average since 1900 is 16.46x. The fathers of security analysis, Benjamin Graham and David Dodd, suggested that this is the preferred method of looking at valuations since it tends to smooth out the variations of the business cycle.
The graph below shows the valuation history based on this method. While things certainly look cheaper now than back at the height of the dot.com bubble, they do not look particularly cheap against the broad sweep of history.
However, it does not take the level of interest rates into consideration at all. The graph below is my crude attempt at a synthesis between the entirely interest rate-driven Fed Model, and the trailing 10-year earnings approach.
The former suggests that stocks are undervalued to an extent that strains credibility (if the earnings yield were to match the current 10-year T-Note, we would be talking about a P/E of 52.1x, and the “fair value” of the S&P 500 would be 4,975). The 10-year trailing P/E method suggests that the S&P 500 is overvalued by 18.2%.
The graph below shows the ratio of the earnings yield to the 10-year T-note (at year end, except for the last observation, which is current) since 1900. The Fed Model suggests that any reading under 1.0 means that stocks are overvalued, and above 1.0 as undervalued (keeping in mind that since it is a ratio, a reading of 0.50 would be equivalent in overvaluation to a reading of 2.0 in undervaluation).
Clearly we are not at record levels of undervaluation by this synthesis approach, but stocks do look much better than they have at any point since the mid-1950’s (with the exception of the end of 2008). There does seem to have been a sea change in valuations that happened in the 1950’s. If only the period since 1960 is considered, we are almost right at the historical average valuation based on the trailing 10-year P/E ratio.
Long-term investors should start to take advantage of the current valuations. However, I would not be shooting for the stars. Look for those companies with solid dividends (say over 2.5%), low payout ratios, solid balance sheets, and a history of rising dividends, which are still seeing analysts raise their estimates for 2012, or are at least not cutting them aggressively. I don’t know if you will be happy doing so next week or even next month, but I am pretty sure that you will be quite satisfied five years from now if you do so.
The graphs above cover a very long time period, and the under- and overvaluations suggested by it can persist for many years, so this is not in any way a short- or even medium-term timing tool. The results it suggest though are in line with what a simple comparison dividend yields versus treasury interest rates would suggest.
It is also entirely possible (likely?) that most of the gap in valuation (since 1960) is closed by interest rates heading up rather than by P/Es expanding, or that the expansion we do see in P/E ratios comes from the wrong source — falling earnings rather than from higher prices.
The analysts who track the individual companies are still looking for solid growth in earnings next year, so unless we see the current trend towards cutting estimates continue or even accelerate, it is unlikely that the gap gets closed through falling earnings alone. The graph above might well be making a better case against investing in long-term government bonds than it is in making the case for investing in stocks.
From the point of view of the long-term investor, this still looks like one of the best times to invest in my lifetime.
Zacks Investment Research
