A Cautionary Note About Robert Shiller’s CAPE

by Stephen E. Wilcox

One measure of market valuation that has become widely popular is the CAPE, the cyclically adjusted price-earnings ratio. It was developed by Robert Shiller, a professor of economics at Yale University and author of “Irrational Exuberance” (Crown Business, 2006).

Shiller defines the numerator of the CAPE as the real (“inflation-adjusted”) price level of the S&P 500 index and the denominator as the moving average of the preceding 10 years of S&P 500 real reported earnings, where the U.S. Consumer Price Index (CPI) is used to adjust for inflation. The purpose of averaging 10 years of real reported earnings is to control for business cycle effects. Table 1 provides a detailed example of how the CAPE is determined.

It is frequently suggested that the CAPE should be considered a mean-reverting series. As can be seen in Figure 1, the CAPE was reported as 23.35 during the month of July 2011 on the Irrational Exuberance website (www.irrationalexuberance.com). Per an analysis frequently used in practice, comparing the July 2011 CAPE to its long-term average of 16.41 indicates that U.S. stocks are currently overvalued by 42.3%.

...To continue reading this article you must be registered with AAII.

Gain exclusive access to this article and all of the member benefits and investment education AAII offers.
JOIN TODAY for just $29.
Register for FREE
to read this article and receive access to future AAII.com articles.

Log in
Already registered with AAII? Login to read the rest of this article.
Stephen E. Wilcox is a professor of finance at Minnesota State University, Mankato.


Edward from Pennsylvania posted over 2 years ago:

Another "This time is different" argument. How about corporate earnings are currently based on record profit margins that will surely revert to the mean. Look out below stock market.

G from Massachusetts posted over 2 years ago:

This is an excellent analysis. Thankyou.

G from Massachusetts posted over 2 years ago:

Edward, I agree margins are very high and will mean revert however I do not think the author was trying to discredit Schiller or say this time is different but merely pointing out some flaws in the data used to arrive at his conclusion that the market is overvalued. The Fed Model would indicate the fair value P/E on the S&P 500 with the 10 year at 2.22% would be 35. Because rates are so artificially depressed I don't give that model any credence. All models have flaws and you need to understand what you are measuring and how it is being measured if you are using it to make decisions in your portfolio.

C from Rhode Island posted over 2 years ago:

Thank you Stephen for a well thought out critique of Shiller's CAPE-based spy valuation. Lots of excellent points raised that are worthy of consideration.

What your post here does not assess though is the robust relationship between CAPE-P-E based valuation and subsequent 10 year returns
on spy as espoused by John Hussman.


The point here is that all models are 'wrong' but need to be assessed on their usefulness for intended purposes.

You have shown CAPE-based P-E model to be 'wrong', John Hussman has shown it to be useful regardless.

Jim from Florida posted over 2 years ago:

OK, I grant that the CAPE may have a few "minor" flaws. But they are in fact minor. The only one quantified here is the business cycle and that leads to the "corrected" conclusion that the market is 34.7% overvalued compared to 42.3%. So what, the fundamental conclusion that the market is significantly overvalued still stands.

The author does not address the larger question of whether the "bubble" period from 1995 should be included in the computation of the long-term averages. Excluding that atypical period leads to a long-term average of the CAPE in the area of 14 making the current market even more overvalued relative to this norm.

Shiller's CAPE is a breath of fresh air compared to the usual valuation arguments heard on CNBC or read in the media that incorrectly conclude that stocks are a bargain.

John from Florida posted over 2 years ago:

This is a very thoughtful article. Something not discussed is which companies are included in this study of PE's. The constantly changing companies included in the S&P 500 make this 10-year smoothing PE not a good barometer of "today's" PE. How many companies were in the S&P in 2001 and are no longer in the S&P today? I think this is a good barometer to think about valuations but I would weigh trailing and consensus forward PE more heavily than this measurement. Just as we should take changes of accounting standards into account when reviewing the PE ratios, we need to take the composure of the index into account when deciding how relevant the measure is to today’s earnings reports for the companies in the index.

David Merkel from Maryland posted over 2 years ago:

You should probably revisit your thesis. I used to think that the CAPE10 was garbage, but after I did this analysis, I changed my mind:


Underlying value in the market changes slowly, though prices may change rapidly for a variety of reasons. Thus we need longer-dated measures of valuation to filter the noise. This would include:

CAPE10, and
the Q-ratio.

The market does not do well when these ratios are high, and tends to do well when they are low.

Corporate profit margins are high now, but that series tends to mean revert. Current and forecast earnings make the market look cheap, but who can tell how long that will last? High profit margins invite competition, and quickly.

Elio from Illinois posted over 2 years ago:

"However, problems can arise if the method used to measure inflation changes over time."


Price/Earning= P/E

Price/CPI= Real Price
Earning/CPI= Real Earning

Real Price / Real Earnings = Real P/E

Price/CPI /(divided) Earning/CPI =
Price/CPI *(multiplied) CPI/Earning

Therefore the two CPIs cancel out and we have


Which says that inflation never get into the formula of the CAPE.

Surprise suprise!!

Elio from Illinois posted over 2 years ago:

"However, problems can arise if the method used to measure inflation changes over time."

Price/Earning= P/E;
Price/CPI= Real Price;
Earning/CPI= Real Earning;
Real Price / Real Earnings = Real P/E;
But Price/CPI /(divided) Earning/CPI = Price/CPI *(multiplied) CPI/Earning; Therefore the two CPIs cancel out and we have Price/Earning=P/E; Which says that inflation never get into the formula of the CAPE. Surprise suprise!!

Steve from Minnesota posted over 2 years ago:

Hello, all. I am the author of this article. I appreciate your comments and interest. My responses to some of your comments follow:

Edward (and others): Note the article does not say anything about me being bullish or bearish. The point made is that the popular analysis that compares the CAPE to its long-term average currently produces an overly- bearish estimate of the worth of U.S. equities. I don't think U.S. equities are overvalued by 42.3%.

G: Loved your second post: "All models have flaws." Yes, they do.

C: Yes, like all valuation indicators with price in the numerator, "high" is bad and "low" is good. I've seen those 10-year rolling return graphs you mention and they are certainly visually appealing. But if you put CAPE as the explanatory variable in a regression with 10-year real returns as the dependent variable you would see there is a huge standard error on the coefficient estimate. Also, the point is that the July 2011 CAPE is artificially (for lack of a better word) high for reasons mentioned in the article.

J: The reason that the business cycle problem was the only one quantified is that one needs to make assumptions to quantify the other two problems. The difference in valuation can be quite significant, however, depending on the assumption. Here is a link where a higher estimate of CPI inflation is used: http://www.saturna.com/navigator/20100802.shtml#

John: Good point. The composition of the S&P 500 has changed considerably over time and that could possibly create comparison problems for the CAPE time series.

David: I certainly don't think the CAPE is "garbage." As I stated in an earlier comment, "low" is better than "high" for valuation ratios and the CAPE is no different in that regard. What bothers me is the analysis that compares the CAPE to its long-term average. That is a flawed approach which currently provides an overly bearish view of the worth of U.S. equities. However, you can't rule out some future point in time where that sort of analysis would provide an overly bullish estimate for U.S. stocks. My point isn't that CAPE is flawed; rather, the point is that any time series analysis of CAPE has some serious limitations.

Elio: You and I have exchanged e-mails and I hope I have answered your questions. If not, feel free to contact me again.

I would like you all to know that I hold Shiller's entire body of work in highest regard and think he deserves a Nobel prize.

Best wishes to you all, SW

D. from Maryland posted over 2 years ago:

You mentioned SFAS 157 Fair Value Measurements & “available for sale” investment securities.

Isn't there another standard that handles things NOT available for sale? How much "discretion" do banks have to shift losses into the "not for sale" category, and how does that distort the value shown on the books?

Stephen from Minnesota posted over 2 years ago:

Hi, D. My understanding is "available for sale" securities are those securities an institution intends to trade and they must be marked-to-market. The securities an institution plans to hold until maturity do not fall into this category. I have seen research that suggests that banks did use accounting discretion during the financial crisis to protect their capital positions. However, I am not an accountant nor do I personally do research in this area. If you are interested in this topic, I'm sure Google or JSTOR could be used to find some current references. I do know that the credit spreads on ABS and MBS have come down significantly post-crisis which suggests some of the apparent losses during the crisis might have been illusory. But you better check this out yourself. Good luck to you, SW

Dominick from Washington posted over 2 years ago:

Interesting article and analysis on the CAPE. I personally am not too knowledgeable in this subject but nonetheless find it beneficial and intriguing.

One question I have though is would it possibly be more accurate to measure the CAPE based on the median 10 yr real earnings instead of average 10 yr trailing earnings? You mentioned that the current CAPE reflects the negative effects of the most recent crash and the dot-com bubble, dragging the overall average significantly down. Perhaps a median figure instead of moving average could eliminate the effects of those outliers. Just a thought.


Stephen from Minnesota posted over 2 years ago:

I'm not sure how well that would work, Dom. I understand your reasoning that choosing the median might avoid some of the problems associated with the mean, such as the problem caused by the large security losses that occured in a single year, 2008. But, from my perspective, the larger problem is with any time series analysis (comparing the present to the past) of CAPE. In my opinion, there have been too many changes to business practices, accounting standards, corporate taxation, and inflation reporting to believe comparisons over a 140+ years are truly valid. Take care, SW

James from Ohio posted over 2 years ago:

I read the article. It was extremely well researched and written. I can't quibble with anything that Professor Wilcox said.

I also read the prior posted comments. They are detailed and well-thought out.

However, it all strikes me as "much-ado-about-nothing"

I downloaded the Shiller Excel file that has data back to 1871. I plotted the conventioinal PE Ratio and CAPE. For the 140 years, they almost always moved up and down together. As a natter if fact, their correlation is higher than 75%. They commonly peaked and bottomed at the same time. Sometimes the conventional PE was higher than the CAPE. Sometimes vice versa.

I feel that fundamental analysis and technical analysis techniques are only valuble to the extent that they help us improve our timing as to when to buy and sell in order to increase our investing profits.

With all do respect to Mr Shiller and Professor Wilcox, I, for one, don't see how the technical details of the differences between the conventional PE and CAPE or the weakness of either can help us do that.

If you do, please speak up.

Stephen from Minnesota posted over 2 years ago:

Hi, James. Thanks for your kind words about my article. I'm quite sure you are correct in stating that changes in conventional P/E and the CAPE are highly correlated. However, the point made in my article was that the popular analysis that compares the CAPE to its long-term average is currently (July 2011)providing an overly bearish assessment of the worth of U.S. equities.

Here is a link to a Forbes article that briefly mentions my paper:


As you can see, there is a graph of traditional P/E back to 1960. Note that it currently is significantly under its average since 1960 while (see my paper) the CAPE is currently significantly over its long-term average. Thus, the two measures are giving conflicting views of the worth of U.S. equities. Something is screwed up and I think it is the CAPE analysis. Take care, SW

Jason from Ohio posted about 1 year ago:

Great article, but I think it misses the point that today's earnings, the denominator, are blow-the-doors-off great due to productivitiy enhancements. Some of those higher margins may be sustainable, but it is more likely that some of the cost savings will be given back.

You need to log in as a registered AAII user before commenting.
Create an account

Log In