In the wake of the financial crisis, I wrote a fair number of articles on market level, bullish in nature, based on considerations such as GDP/S&P 500 (SPY), Corporate Profits/S&P 500, margins, interest rates, financial stress, etc. These articles were generally accurate in their outlook, and relied on reversion to the mean.
By February this year, I became concerned that the market was above its midpoint, so that logically there was no reason to expect above average long-term returns. Investing with that in mind has proven difficult. Since then, I've been trying to come up with a logical way of thinking about investing in a market that's above its midpoint, in the context of historically low interest rates. Reversion to the mean, as I was using it, misses the point: the market often overshoots and just keeps on going. This article presents my current thinking.
I arrive at normalized earnings for the S&P 500 by multiplying current revenue by historical average margins. As reported earnings for 2013 are expected to come in at $97.83, with a profit margin of 8.6%. Adjusting that back to a ten-year average margin of 6.35%, normalized earnings will be $72.23.
Long-Term Growth Estimates
To come up with growth figures, I consult the SEP (Summary of Economic Projections) that comes with the FOMC announcement. Long-term inflation is expected to come in at 2%, and real growth in GDP to settle around 2.3%. Adding the two estimates, nominal growth going forward would be 4.3%. S&P 500 companies, in the aggregate, should be able to achieve that level of growth. Here's what I look at:
Using a DCF calculator to do the math, I input $72.23 as earnings, to increase 4.3% ad infinitum, and specify that the investor wants 9% return on investment. I did an article, asking whether that was a realistic expectation, and it continues to draw page views, over a year from its publication. Investors like that 9%.
Index Level Consistent with Assumptions
The above assumptions produce an indication of 1,600 for the S&P 500. To justify the current level of 1,800, the investor's expected returns would need to be reduced to 8.5%. That's feasible: the 9% was nice, but with 10-year treasuries yielding 2.76%, maybe they should be willing to settle for 50 basis points less.
The Slippery Slope
Suppose, with the 10-year as low as it is, and with market returns as high as they have been, investors begin to reduce their long-term required return. Here's how that changes the resulting index level:
- 9.0% - 1,600
- 8.5% - 1,800
- 8.0% - 2,000
- 7.5% - 2,350
- 7.0% - 2,800
With the ten-year at 2.76%, let's add a 5% risk premium, for a required rate of return of 7.76%. Using the assumptions listed above, that would support the S&P at 2,200.
But using normalized earnings of $72.23 deals with much of the risk. After all, as reported earnings will be $97.83 as of the end of the year. If you're willing to shave that risk premium a bit, a 7% rate of return doesn't seem all that low.
Donald van Deventer mentions that the 10-year treasury is expected to reach a 4.743% yield in ten years. Checking data provided by Portfolio123, the average risk premium since 1999 is 2.12%. Adding the two, the required rate of return would be 6.86%, implying 2,900 for the S&P 500.
CAPE and Normalized Earnings
It should be noted that Shiller's 10-year average EPS is another version of normalized earnings, slightly lower for the fact that it doesn't consider real growth. A study of his irrational exuberance spreadsheet reveals that a CAPE of 40 or better was (and is) a very bad level at which to invest in US equities.
The thing is, if investors will settle for 7% expected returns on the S&P 500, the multiple on normalized earnings as computed for the purposes of this article would be 38.6. So the slippery slope can lead us toward another generational high, and more heartache, in due course.
What's Different This Time
In December 1999, when CAPE hit 44.2, the ten-year was yielding 6.28%. That's a far cry from today's 2.76%. 2% inflation is very tame, compared to what many of us lived with in the 1980s.
If the investing public ever comes to accept the current low inflation and long-term interest rates as being the new normal, and likely to persist indefinitely, it will create considerable pressure pushing stock prices higher.
What if interest rates don't revert to the mean?
How Long Can This Go On?
GMO's Jeremy Grantham recently delivered a fine Jeremiad, fulminating against cheap money and culminating in capitulation. He guesses that the market could go 20% to 30% higher, over one or two years, and cautions readers to bear in mind that it will end badly:
My personal guess is that the U.S. market, especially the non-blue chips, will work its way higher, perhaps by 20% to 30% in the next year or, more likely, two years, with the rest of the world including emerging market equities covering even more ground in at least a partial catch-up. And then we will have the third in the series of serious market busts since 1999 and presumably Greenspan, Bernanke, Yellen, et al. will rest happy, for surely they must expect something like this outcome given their experience.
It will go on until something throws it off the rails. That something will most likely be a deteriorated base of financial assets, specifically commercial loans that are being made to low standards, covenant-lite, and packaged for sale. When enough of that builds up, credit markets will seize up, financial stress will increase, and the air will come out of the balloon, rapidly.
The Reluctant Bull
It's a market of stocks. For an individual stock, normalized earnings can be approximated by computing PE5 (5-year average EPS). The metric can be made forward looking by including an estimate for one year into the future. Analyst consensus can be used, or the investor can develop his own estimates.
When looking at PE5 multiples, I prefer to buy at less than 15, but realistically will settle for less than 20 for a high quality company. At 25 I become a seller. Lower multiples are in order for cyclicals or lower quality.
Using the estimates and reasoning developed above, a 9% required rate of return would imply a PE5 of 22. So, erring on the side of caution, 20 seems about right. I've used it as a rule of thumb for many years, with acceptable results. It's also possible to develop an average PE5 multiple for a given stock, and make buy/sell decisions on that basis.
Companies with a history of dividend increases outperformed over the course of the financial crisis. Larry Fink gave a speech in which he asserted that US companies have gone light on capex, and heavy on dividends and buybacks. Combining the two considerations, companies that have the strength to increase dividends and maintain capex above depreciation can still be found at reasonable valuations.
Options are attractive as a source of leverage, or a method of low cost speculation. However, they can also be used to reduce risk, and that's how I'm using them under the present conditions.
To conclude, I would resist the idea of trying to skate along on thin ice and head for shore just before it gives in. I've tried hiding out in low beta cases, but it hasn't been constructive, for the fact that many of them have become over-valued. As long as it's possible to select individual stocks that meet common sense metrics and quality criteria, I intend to participate in the market, while reserving 1/7 of my liquid assets in cash or other relatively stable value investments.
Additional disclosure: I maintain a long-term position in the Vanguard S&P 500 index fund.