2013 has officially begun and we decided its time to do an update on market valuations for the S&P 500. We have some global market valuation data, which we posted earlier. Readers can also check out the latest data for CAPE across the 32 largest equity markets. We hope to post updates on market valuations levels each quarter. Data from our last update can be found here.
The S&P 500 (S&P Indices:.INX) closed at 1,414 for 2012, for a total return of 16%. The S&P 500 (S&P Indices:.INX) seems to be slightly more over-valued today than it was in the beginning of 2012. The market seemed overvalued at the beginning of 2012, but provided a nice return nonetheless. It is important to remember that just because the market is over or undervalued does not mean that future returns will be high or low. The market can stay irrational for months or even years. Additionally, regardless of the overall market valuation, individual value stocks can be found which will outperform (or underperform the market). However, we think the metrics are useful or at the least, interesting. We use six common metrics to arrive at the S&P 5oo ‘ valuation ‘.
The S&P 500 (S&P Indices:.INX) has a PE ratio of 15.96.
Max:123.79 (May 2009)
Based on the PE TTM the market is basically fairly valued. However, we do not think this is a good way of valuing the market. One example can be seen above, the max (of recorded data is May 2009), which was only a few months from the market low of March 2009. To get an accurate picture of whether the market is fair valued based on price earnings ratio, using several years of earnings seems to be a better indicator.
Numbers from Previous Market Lows:
Shiller PE (or CAPE): 22.53
The Shiller PE, invented by the famous Robert Shiller, is calculated using the four steps below:
- Look at the yearly earning of the S&P 500 for each of the past ten years.
- Adjust these earnings for inflation, using the CPI (ie: quote each earnings figure in 2011 dollars)
- Average these values (ie: add them up and divide by ten), giving us e10.
- Then take the current Price of the S&P 500 and divide by e10.
Using ten years of earnings removes the problem we mentioned above regarding PE TTM. CAPE is extremely high currently, using the mean or median.
The most common criticism of Shiller’s method is that it includes years like 2008 and 2009 when earnings were severely depressed. However, we believe that argument shows why the Shiller PE is suprieror to the PE TTM. Furthermore, earnings from depressed years usually are balanced out by earnings from bubble years. In this case, earnings were high due to the housing bubble from 2003-2007. Furthermore, the Shiller PE (with the notable exception of the mid 1990s to 2000) has been a much better indicator of market bottoms and tops than PE TTM. Two recent examples prove this point. In March 2009, the Shiller PE was at 13, while PE TTM was at 110. At the market top in October 2007, the Shiller PE was at an extremely high level of 27+, while the PE TTM was only slightly above 20.
In our opinion, The AAII has put out the best criticism of the Shiller PE.
Below are their main points:
- Following the Graham-Dodd recommendation, Shiller uses a 10-year moving average of earnings in computing the CAPE. According to data compiled by the National Bureau of Economic Research, economic contractions have become shorter and expansions longer in recent years. Measured peak to peak, the average is five years and six months.
- In determining the CAPE (Shiller PE), reported earnings are adjusted for inflation using the Consumer Price Index, where real values reflect current-period purchasing power. However Shiller uses data going back to 1871, when inflation was measured differently than today. There is an approximate 7% difference in the annual CPI inflation rate based on methodological changes over the last couple of decades.
- Tax codes have changed considerably since 1871. You can read further discussion of this issue above.
Those points are valid. Additionally, we do believe that Shiller takes a bit too much of a conservative approach. For example, using 5-7 years of earnings may be a better indicator than 10 years (would be great if someone could back-test that). Nonetheless, while Shiller’s method is far from perfect, we think it is still valid and the best method of broad market valuation for the reasons stated below:
- The Shiller PE shows markets overvalued by approximately 40% (at time of the article) and using a six year business cycle (as suggested by the AAII) the market is still overvalued by nearly that amount. This could also be the case historically (again would be great if someone would backtest), but all we have is the evidence in front of us. Therefore, five versus ten years is not a reason to completely disregard the Shiller PE.
- Differences in inflation do impact the Shiller PE, but it would likely be small, especially if one decreases the time period in which earnings are measured (as the AAII suggests). Furthermore, there are so many ways to calculate inflation that the point becomes moot (or somewhat of a political debate) . Finally, any changes to inflation will apply to any earnings metric such as ( PE TTM, six-year P/E etc.), so it works
- Same point with accounting changes. Accounting changes would also impact P/B or ANY valuation metric that is used, so this criticism should not be limited to CAPE.
- Finally, go with what works! CAPE has proved itself to be a good market valuation indicator. Indeed, Robert Shiller stated on February 22nd, 2009, that based on his methodology, equities were cheap for the first time in many years. That was only two weeks before one of the biggest bull market rallies in history. Other historic data backs up CAPE.
CAPE data from prior market bottoms:
Mar 2009 13.32
Mar 2003 21.32
Oct 1990 14.82
Aug 1982 6.64
Oct 1974 8.29
Oct 1966 18.83
Oct 1957 14.15
June 1949 9.07
April 1942 8.54
Mar 1938 12.38
Feb 1933 7.83
July 1932 5.84
Aug 1921 5.16
Dec 1917 6.41
Oct 1914 10.61
Nov 1907 10.59
Nov 1903 16.04
In conclusion, we would argue that despite the flaws, CAPE is a good valuation metric.
Robert Shiller stated in an interview that he believes the S&P500 will be at 1430 in 2020. Today, the S&P 500 is above that level at 1,466. What will the market return? That is impossible to predict, as even Shiller noted in the interview. He was giving a best guess estimate. However, assuming that the ratio reverts to the mean and there is normal historic earnings growth, real returns should be barely positive.
Our colleague, Rob Bennett has created a calculator to predict returns using CAPE. His market return calculator shows 2.20% real returns over the next ten years. He explains his methodology further on his site, which can be found through the above link.
|Stock Market||Best Possible||Lucky||Most Likely||Unlucky||Worst Possible|
|10-Year Percentage Returns||8.20||5.20||2.20||-0.80||-3.80|
|20-Year Percentage Returns||7.35||5.35||3.35||1.35||-0.65|
|30-Year Percentage Returns||8.20||7.20||6.20||5.20||4.20|
|40-Year Percentage Returns||7.17||6.27||5.37||4.37||3.37|
|50-Year Percentage Returns||7.20||6.40||5.60||4.90||4.20|
|60-Year Percentage Returns||7.65||7.00||6.35||5.75||5.15|
Current Price over book value:2.32
The number for price over book comes from Barron’s.
Price over book is not commonly used to measure the S&P 500. However, we thought that the metric would be great for measuring market valuation, because as Tweedy Browne, David Dreman and others demonstrate through extensive research, a basket of the lowest P/B stocks dramatically outperform baskets of higher P/B stocks (and the market overall). Logic would state that the P/B metric would be useful for evaluating the market as a whol.
In regard topic data there is very little on P/B, besides the average of 2.41 over the past 30 years. In fact, I have personally seen historic P/B data on the Russian stock market. For some reason the sell-side (and others) have done little to no research on this topic.
Book value is considered a better measure of valuation than earnings by many value investors,especially Marty Whitman. He notes that book value is harder to fudge than earnings. In addition, book value is less impacted by economic cycles than earnings. P/BV therefore provides a longer term accurate picture of a company’s value, than a TTM P/E. The market is slightly undervalued based on P/B.
Current Dividend Yield 2.13
Min:1.11% (Aug 2000)
Max:13.84% (Jun 1932)
Is the stock market cheap based on dividend yield? It depends on how you look at it. Many investors compare dividend yield to the risk free rate (10 year Treasuries), in which case stocks are cheap. However, just looking at dividend yield alone, would indicate equities should decline or dividends have to increase quickly.
Min:1.53% (Jul 2012)
Max:15.32% (Sep 1981)
Dividend yield data from previous lows:
Mar 2009 3.60
Mar 2003 1.92
Oct 1990 3.88
Aug 1982 6.24
Oct 1974 5.17
Oct 1966 3.73
Oct 1957 4.29
Jun 1949 7.30
Apr 1942 8.67
Mar 1938 7.57
Feb 1933 7.84
July 1932 12.57
Aug 1921 7.44
Dec 1917 10.15
Oct 1914 5.60
Nov 1907 7.04
Nov 1903 5.57
Data and chart courtesy of [multpl.com]
Current Tobin’s Q 0.94
As can be seen from the above charts, the market is over-valued based on tobins Q.
What is Tobins Q?
Here is a brief explanation from Investopedia:
A ratio devised by James Tobin of Yale University, Nobel laureate in economics, who hypothesized that the combined market value of all the companies on the stock market should be about equal to their replacement costs. The Q ratio is calculated as the market value of a company divided by the replacement value of the firm’s assets:
Investopedia explains Q Ratio (Tobin’s Q Ratio)
For example, a low Q (between 0 and 1) means that the cost to replace a firm’s assets is greater than the value of its stock. This implies that the stock is undervalued. Conversely, a high Q (greater than 1) implies that a firm’s stock is more expensive than the replacement cost of its assets, which implies that the stock is overvalued. This measure of stock valuation is the driving factor behind investment decisions in Tobin’s model.
As can be seen from the above charts, the market is over-valued based on tobins Q. The data comes from Doug Short. This is the most accurate data that is available. It is impossible for the data to be 100% precise because the Federal Reserve releases data related to Tobin’s Q on a quarterly basis. The best that can be done is to extrapolate the data and try to provide the most accurate data possible based on the change in the Willshire 5000. This is what we and Doud did to get the current number. This method has proven extremely accurate for calculating Tobins Q on any given day.
The current level of 0.94 compares with the Tobin’s Q’s average over several decades of data of approximately .72. This would indicate that the market is overvalued.
In the past Tobin’s Q has been a good indicator of future market movements. In 1920 the number was at a low of .30; the next nine years were some of the best ever for (long) investors. In 2000, Tobin’s Q almost reached a record high of nearly 2, and the market declined subsequently nearly 50% within a few years.
Tobin’s Q at prior market bottoms. Data courtesy of Doug of dshort.com:
Market High 1929: 1.06 (This is not the highest number ever reached, rather the peak before the 1929 crash.)
One reader pointed out the following interesting point about Tobin’s Q:
The idea behind Tobin’s Q is interesting but a modern complexity relates to the amount of money spent on R&D. It seems clear to me that the investment in many products is in proprietary technology and not in manufacturing facilities. Look at Amgen, Apple, Intel or Microsoft. Fabless Semi’s for instance. Most valuation techniques would give you a way to adjust accounting values for the R&D. I really think Tobin’s Q would be more valuable if the amount of research that had been accomplished in the last decade could be included. It may then provide a better valuation measure for modern times. A good rule of thumb may taken from Intel and how it bottoms at 2 x book. Maybe the value of its R&D is 2 x book or for for the market, 2 x Tobin’s Q. I don’t think Tobin’s Q adjusts for research although I haven’t read the techniques in a while.
Finally, we use one metric which is slightly less Scientific. Market sentiment can be a good way to gauge valuations. Usually when the retail investor is bullish it is a contrarian sign of a peak, and when they are super bearish it is a sign of a coming bull market. Using investor sentiment from the AAII has proven these points. In 2009, 70% of investors were bearish, the market is up approximately 130% since that time. More on the methodology below:
According to the AAII, “the AAII Investor Sentiment Survey measures the percentage of individual investors who are bullish, bearish, and neutral on the stock market for the next six months; individuals are polled from the ranks of the AAII membership on a weekly basis. Only one vote per member is accepted in each weekly voting period.”
AAII sentiment survey data from 1/02/2013 AAII Sentiment Survey:
Investors are a bit more bearish than usual, which is a good sign for long investors.
Sentiment Survey Past Results
For all historic data on the AAII survey back to 1987 click on the following link:http://www.aaii.com/sentimentsurvey/sent_results.
Below some recent data and historic data to prove our point.
|Table 1. Historical Averages & Extremes of AAII Member Sentiment*|
|[i]*data covers period from July 24, 1987 to September 2, 2004.[/i]|
We see the historical extremes for bullish, bearish, and neutral sentiment. Bullish sentiment reached its highest levels on Jan. 6, 2000 — the height of the tech bubble — at 75.0%. Levels of extreme bullishness — if viewed as a contrarian indicator — would lead one to believe that a market decline is in the wings, and vice versa.
Bullish Neutral Bearish
Max 75.00% 62.00% 70.00%
March 2009 18.92% 10.81% 70.27%
March 2003 34.3% 14.30% 51.40%
Oct. 1990 13.00% 20.00% 67.00%
1. P/E (TTM) – Fairly valued- 15.96
2. P/E 10 year – overvalued 22.6 overvalued
3. P/BV –Slightly undervalued
4. Dividend Yield 2.13%– Indeterminate: Overvalued on a historic basis, but very attractive compared to treasuries: 1.47%
5. Tobin’s Q – Overvalued 0.94
6. AAII Sentiment- Slightly Undervalued