Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.
Support MFO
Donate through PayPal
Shiller’s Powerful Market Indicator Is Sending A False Signal About Stocks This Time
I had trouble accessing Ted's link, but I did read the article since we get the WSJ daily.
Shiller's CAPE model is really not a powerful signal that forecasts market direction. That's not me talking. A comprehensive study was completed by Vanguard a few years ago that explored the potential of CAPE and other candidates for that purpose. It is a good study. Here is a Link to its findings:
The Vanguard study postulated a dozen or so candidate signals that might forecast market direction. None reached the 50% reliability score. Shiller's CAPE model was the best of those examined in the study.
But it failed more often than it succeeded. The monthly and the annual actual data had a huge cloud scatter around his forecast correlation. Measures of the correlation accuracy are disappointingly low. The CAPE as a market signal is very weak and not trustworthy.
Shiller is an honest researcher and near the end of the referenced WSJ article he cautions readers about the limitations of CAPE. As always, investor beware. There are no magic elixirs that project market returns, especially since the market itself is constantly changing and evolving.
Fascinating. Even more interesting for me is that for the last 4y it is extremely hard to find anything with consistency that has outperformed the etn CAPE (not SCHD, DVY, insofar as divs correlate w/ value; not RPV; not NOBL and OUSA more recently; not the good active value funds like ACIIX). Only SPHD, again recently, and again not really the right space either.
The crucial point of the article questions the current suitability of some of the data input to CAPE. Changes in accounting rules have "made the CAPE artificially high relative to its history."
CAPE uses generally accepted accounting principles (GAAP) as an input, but this varies over long time periods because of GAAP revisions. From the article:
"An alternative CAPE, constructed by The Wall Street Journal, uses the same methodology as Mr. Shiller, but relies on a more consistent earnings measure: The Commerce Department’s quarterly data on total U.S. after-tax corporate profits. Then for prices, it uses Federal Reserve data on the total value of the U.S. stocks, rather than the value of the S&P 500. The result: Stocks look much cheaper than Mr. Shiller’s data suggests."
The alternative input data is "based on corporate tax and financial data" following "a consistent standard over the decades".
"From the early 1960s through 2008, the two CAPE measures move nearly lockstep with one another. It is only after the financial crisis that the two measures diverge, with the corporate-profits CAPE rising much more slowly than Mr. Shiller’s measure does."
This suggests that changes to GAAP have pushed Mr. Shiller’s CAPE readings artificially higher.
Note: "" indicates text directly from the WSJ article. All text emphasis was added.
I see many of you are also intrigued by the Shiller CAPE work. Initially I gave it more predictive power than it practically warranted. Here is a Link to a paper that explores any relationship in depth:
The first figure shows the correlation curve with the data. It definity challenges my overvaluation of the proposed model. The correlation is underwhelming.
The scatter around the correlation best fit curve demonstrates that the PE is an unreliable signal, even with accurate earnings forecasts. That's disappointing. The r- squared statistic is near zero indicating little value as a forecasting tool, at least for the one-year PE measurements.
Shiller recognizes the shortcomings of his one-year CAPE. He is very humble over his contributions. That's reflected both when he writes and more so when speaks.
@Derf- You are very welcome, sir, and thanks for your note. And yes, I totally agree about things working until they don't. Especially electronic "things" (and also my memory, as of late...).
Hmm... I never knew anyone wanted CAPE to predict market direction. I thought CAPE was a way to identify likelyhood of outperforming stocks, and hence, one invests in it.
the etn CAPE has gotten totally conflated here with the Shiller p/e analysis tool C A P E
happens all the time here, like w dca and etf and all those things which if you do not lowercase them or separate them become trading symbols automatically
well, this year has favored value over growth, and the underlying bonds have likely done well based on the rest of Doubleline's holdings, but it seems to have done well when growth outperformed value, and bonds were under more stress..?
Any insight that DSENX is doing well sofar, YTD is over 15% while the broader S&P 500 is up 7%. When will it revert to mean?
DSENX does seem wierd. It sorta simulates CAPE using futures and stuff I think and then has a bond overlay. I don't think it is a proxy for CAPE at all. It must be using some CAPE principles that's all. As long as it works. I simply bought it as an "alternative" fund.
@jlev and VF, thanks for sharing. DSENX has done well since inception and I try to better understand why and how it works, especially with CAPE. Maybe I am thinking too much.
Read up. I believe it is CAPE, effectively, and then the extra bond sauce accounts for the outperformance. I mean, just track DSENX vs CAPE.
Three years old now.
Every month it buys low and sells high SP500 per C A p/e. Study it. What would mean reversion look like given this method?
It is now over half of my holdings.
What is particularly interesting is that the bond sauce is additive / augmentative, whereas if you held CAPE 80% and (say) PONDX 20%, the total result is notably less than CAPE alone. I would like to see an explan for that.
Instead of trying to devine an absolute level of CAPE in order to make a determination of market overvaluation, the measurement of excursions of CAPE levels in standard deviations above it's long term "mean" with subsequent application of a well tested tactical strategy , has produced risk adjusted alpha vs. buy & hold, on average, since 1927. https://docs.google.com/presentation/d/1mdon_cto48rvs2_lKWyMWrfqSIh8K0phfe7tThle8qQ/edit?usp=sharing
It can, on occasion, be prudent to employ tactical strategies in order to preserve capital from ravages of market volatility created during high valuation periods.
The WSJ monthly rankings of funds has DSENX and CAPE at #1 and #2 for one-year performance in the Large Value category. CAPE still has only about $35M in assets, so come on, momentum guys!
Glad to see MOAT up there as I got back into the fund after it had a terrible patch during the energy price drop. The fund has changed its methodology, so it's unlikely to wind up again overweight one sector.
Comments
I had trouble accessing Ted's link, but I did read the article since we get the WSJ daily.
Shiller's CAPE model is really not a powerful signal that forecasts market direction. That's not me talking. A comprehensive study was completed by Vanguard a few years ago that explored the potential of CAPE and other candidates for that purpose. It is a good study. Here is a Link to its findings:
https://personal.vanguard.com/pdf/s338.pdf
The Vanguard study postulated a dozen or so candidate signals that might forecast market direction. None reached the 50% reliability score. Shiller's CAPE model was the best of those examined in the study.
But it failed more often than it succeeded. The monthly and the annual actual data had a huge cloud scatter around his forecast correlation. Measures of the correlation accuracy are disappointingly low. The CAPE as a market signal is very weak and not trustworthy.
Shiller is an honest researcher and near the end of the referenced WSJ article he cautions readers about the limitations of CAPE. As always, investor beware. There are no magic elixirs that project market returns, especially since the market itself is constantly changing and evolving.
Best Wishes.
Derf
CAPE uses generally accepted accounting principles (GAAP) as an input, but this varies over long time periods because of GAAP revisions. From the article:
"An alternative CAPE, constructed by The Wall Street Journal, uses the same methodology as Mr. Shiller, but relies on a more consistent earnings measure: The Commerce Department’s quarterly data on total U.S. after-tax corporate profits. Then for prices, it uses Federal Reserve data on the total value of the U.S. stocks, rather than the value of the S&P 500. The result: Stocks look much cheaper than Mr. Shiller’s data suggests."
The alternative input data is "based on corporate tax and financial data" following "a consistent standard over the decades".
"From the early 1960s through 2008, the two CAPE measures move nearly lockstep with one another. It is only after the financial crisis that the two measures diverge, with the corporate-profits CAPE rising much more slowly than Mr. Shiller’s measure does."
This suggests that changes to GAAP have pushed Mr. Shiller’s CAPE readings artificially higher.
Note: "" indicates text directly from the WSJ article. All text emphasis was added.
I see many of you are also intrigued by the Shiller CAPE work. Initially I gave it more predictive power than it practically warranted. Here is a Link to a paper that explores any relationship in depth:
http://seekingalpha.com/article/3987114-predicting-stock-market-returns-using-shiller-cape-pub
The first figure shows the correlation curve with the data. It definity challenges my overvaluation of the proposed model. The correlation is underwhelming.
The scatter around the correlation best fit curve demonstrates that the PE is an unreliable signal, even with accurate earnings forecasts. That's disappointing. The r- squared statistic is near zero indicating little value as a forecasting tool, at least for the one-year PE measurements.
Shiller recognizes the shortcomings of his one-year CAPE. He is very humble over his contributions. That's reflected both when he writes and more so when speaks.
Best Wishes.
To my thinking, things work until they don't !
Derf
OJ
happens all the time here, like w dca and etf and all those things which if you do not lowercase them or separate them become trading symbols automatically
I think...
Three years old now.
Every month it buys low and sells high SP500 per C A p/e. Study it. What would mean reversion look like given this method?
It is now over half of my holdings.
What is particularly interesting is that the bond sauce is additive / augmentative, whereas if you held CAPE 80% and (say) PONDX 20%, the total result is notably less than CAPE alone. I would like to see an explan for that.
https://docs.google.com/presentation/d/1mdon_cto48rvs2_lKWyMWrfqSIh8K0phfe7tThle8qQ/edit?usp=sharing
It can, on occasion, be prudent to employ tactical strategies in order to preserve capital from ravages of market volatility created during high valuation periods.
LC Growth: ROLA and LSGRX
If you subscribe, the following link will take you to the relevant page:
http://online.wsj.com/mdc/public/page/2_3061-mfq16_3_CategoryKings.html#catcore1
Glad to see MOAT up there as I got back into the fund after it had a terrible patch during the energy price drop. The fund has changed its methodology, so it's unlikely to wind up again overweight one sector.