DOES IT MATTER IF STOCKS ARE CHEAP?

There seems to be a never ending debate these days about the “value” of the market.  The bulls cite trailing earnings to show low valuations while the bears generally cite forward earnings or cyclical data as evidence that the market might be expensive.  I’ve never found much value in P/E ratios or other such “value” metrics for that matter.  And honestly, unless you’re Warren Buffett or a hedge fund with access to the Board of Directors or other inside info you’re battling armies of analysts using the same metrics in the same manner.  The odds of you finding a diamond in the rough using P/E ratios or similar valuation metrics is frankly, very low.

And now, an interesting bit of research from Draco Capital Management shows that investors might place far too much emphasis on the “value” of the market.  In fact, their research shows that it’s nearly impossible to decipher future returns using present valuations (via Advisor Perspectives):

“But let me run you through a few charts that may raise some questions about your blind faith in the Price-to-Earnings multiple, the most popular way of measuring value.

What this first chart does is take a snapshot of the market’s P/E ratio on January 1, 1900.  Then it fast-forwards three years to see how much money you made or lost if you made an investment in the Dow.  Those two pieces of data become one little dot on our scatter-plot chart.  I’ve plotted points for every month in the 20th and 21st centuries.”

“In this chart we’re relating present valuations to future returns.  We’re checking to see if buying the stock market when it’s cheap translates into profits down the road.

Over a three year window, the data tell a fairly convincing story.  Valuations matter a lot less than we’re comfortable admitting.  As indicated by our regression line, there is a modest correlation between present valuation and future return, but with an R-squared of 0.03, it should be ignored.  So there you go.

Over a 3-year window, there is no correlation between ttm P/E’s and the degree of future profitability.”

They ran the same data over a longer period.  The results were pretty similar:

Let’s broaden our window a little further.  Let’s widen our scope to 10 years.  Does buying stocks when they are particularly cheap translate into more profitability for the coming decade?

“The visual correlation is a little more clear.  Our R-squared is a bit higher but is it even at the level where it’s statistically meaningful?  I’m still not sure I’m prepared to say that buying at lower multiples translates to higher profitability.

I feel good about saying one thing, though.  There are few guarantees in this business, but — surprise! — if you buy stocks when they are craaazy cheap and can hold onto them for 10 years, the odds suggest you’ll make money.”

Their conclusion:

“1. Correlation between earnings multiples and subsequent returns is loose.  Statistically speaking, it’s almost non-existent.

2. Unless you’re dealing with long holding periods, pretty much any outcome is in play.  Markets that are historically cheap can get a lot cheaper and take a lot longer than most humans are comfortable waiting before swinging back into profitability.

3. In truth, this kind of analysis is only really useful during a few select moments in history.  It’s only helpful at identifying extremes when the probabilities start to skew in the rational investor’s favor.  The problem is that during those extremes is when the fewest number of investors are bothering with this type of analysis.  They’re lost in all the other noise of the moment.  They’re buying tech stocks in 2001 because, it’s the future, maaan!  Or they’re cowering in their bunker in 1948.  Or broke in 1933.  Or high on the Nifty Fifty in 1965, among other interesting things.”

As they say, past results are not indicative of future returns.  Beware of those data mining using the rear view mirror.  These various valuation ratios might make the average financial advisor sound smarter than the average bear, but that doesn’t mean he/she is utilizing information that will necessarily lead to superior returns….

Cullen Roche

Mr. Roche is the Founder of Orcam Financial Group, LLC. Orcam is a financial services firm offering research, private advisory, institutional consulting and educational services.

More Posts - Website

Follow Me:
TwitterLinkedIn

  • Mario

    interesting.

    How is he getting that regression line though?

    If every dot represents a month in the 20th and 21st century to date with either a 3 or 10 year holding period…well it looks to me like statistically the odds are in your favor if you buy a stock under-valued around the $10-$20 range and hold it for either 3 or 10 years. Based upon the x-axis brackets there appears to be a whole lot of returns being made in both time frames wouldn’t you agree…perhaps in the 5-10% rate on average just by eyeballing it?

  • Mario

    also it appears that the more expensive the stock is the less returns you are likely to receive from it…even at under-valued levels. That seems to be both intuitively correct but perhaps even more accurate than our intuition would think would you agree?

  • Vikash

    Is it correct to do such an analysis using price-weighted DJIA?.

    If I remember correctly, Burton Malkiel in A Random Walk Down Wall Street did similar analysis using S&P500 and found a much more stronger relationship with the 10yr holding period.

  • Andrew

    All that matters is whether or not you think there will be someone slower than you or less smart than you to buy/sell the stock at a worse price.

    Hedge funds and investment banks have virtually unlimited leverage to send the market indices to seemingly impossible highs and lows. Don’t ya remember Y2K.

    Obviously those that administer the trading accounts of the biggest dumbos and losers have a rough idea when to call the tops and bottoms.

  • Johnnythunders

    You mention cyclical data – the cyclically-adjusted Shiller CAPE predicits future 10 year returns with an in-sample R2 of 32% from 1881 to date. See prof. Shiller’s website. Looks a little better…

  • b_b

    PE’s are a very basic tool for valuation.

    A company on 20x earnings could be much cheaper than a company on 8x. Reasons may include;
    - Lower premium (higher discount) to book value implying differing earnings potential or break-up value
    - Agency risk (ploughback & strategy)
    - Leverage
    - Capital structure
    - Marginal ROE versus ploughback
    - Industry dynamics (competition / substitution)

    I am not at all surprised at the results.

  • Pod

    Does it (value) matter? Yes, but only if you measure it correctly. Try using a normalized P/En and you will get a very different result. The author is confusing cyclical and secular.

  • Dr. Oliver Strebel

    Cullen Roche: “And honestly, unless you’re Warren Buffett or a hedge fund with access to the Board of Directors or other inside info you’re battling armies of analysts using the same metrics in the same manner.”

    Why should the board of directors tell a hedge fund, pension fund or Warren Buffet the bad news that does not show up in its financial statements? On a visit you will hear just about the glorious plans they have in the future, presented on perfectly styled slides. However, if you visit as a pension fund manager Danone or Nestle, you will leave the company with several suitcases full of chocolat, but thats it ;).

    @the graphs: The question is, if R-squared of a linear fit is a reasonable measure for a relationship between stochastically evolving quantities. I think the interesting data deserve more elaborate analysis ;).

  • Sojourner

    This is nothing but a correlation study, and correlations say nothing about causality. For example, there is an extremely high correlation between the number of ice cream cones that are sold, and drowning deaths. Clear message: eating ice cream causes drownings!! Hardly – just a spurious correlation.

    If you want to talk even remotely about causation, then certainly more variables have to be included. The “regression” line shown in these charts is nothing but “simple regression,” which is a fancy term for correlation. Produce a study using multiple regression analyses and then see what conclusions can be drawn.

  • Max

    12-month PE is a TERRIBLE method because PEs are often low at market peaks, and high at bottoms! The only worse method is forward PE.

  • jt26

    <>

    If you know the # of data points and you can’t answer that question, you have no business talking about statistical analysis …

  • jt26

    Somehow the quote-unquote go sliced out (I’m an html idiot obviously …

    “The visual correlation is a little more clear. Our R-squared is a bit higher but is it even at the level where it’s statistically meaningful? I’m still not sure I’m prepared to say that buying at lower multiples translates to higher profitability.

  • VRB II

    Ok Draco.
    Then what?
    What should one use? U manage money let’s see what u use has done?
    Let’s peel YOUR onion back

  • http://www.butlerphilbrick.com Adam Butler
  • freemarketeer

    I’m under the impression that securities are priced according to expectations of the future. Please explain how the forward multiple is worse than the TTM multiple.

  • SBG

    If you read the whole piece from Draco Capital, Jeffery Dow Jones goes on to state that normalized PE ala Shiller, Hussman, Easterling (I added Easterling) has merit.

    “Over the course of a decade, low normalized PE’s not only imply a greater chance of future profitability, they also suggest a greater degree of profitability.”

    It all comes down to how you calculate your earnings metric.

  • Silalus

    I agree with this philosophically. Clearly understanding and being able to beat mob mentalities in the market (a ton of what we tend to discuss here) trumps trying to beat analysts and fund managers with access to far more sophisticated valuation data.

    But I question the validity of using an R-squared value to prove the point. I’d be more interested in seeing a coefficient of correlation, for example, because the -exact- degree of correlation isn’t very intesting except to produce price targets. Seems like the degree to which you can count on something following up or down -at all- is a lot more important to investment decisions compared to a precise target.

  • In Accounting

    Because analyst consensus is generally overly optimistic. Seems the original source is now behind a subscriber wall so here is the zerohedge coverage:
    http://www.zerohedge.com/article/mckinsey-study-confirms-sellside-analysts-are-conflicted-slow-biased-and-generally-stupid

  • Rich

    The first sentence says it all: “There seems to be a never ending debate these days…” So many things to worry about, so little time to make money. How long are y’all willing to wait for the mean reversion to kick in, if it even exists? 20 years? Potential excess returns range from none to mediocre, depending on the study. And worse, the S&P earnings data seems to be contaminated by revisions, while most of the P/E enthusiasts suffer from look-ahead bias. What was the Dow’s 10 year trailing P/E in 1896? Should you have bought or sold?

    Cullen, I’m with you on the minimal value in valuation metrics. Ya pays your money and ya takes your chances.

  • http://pragmaticcapitalism Michael Schofield

    This is another study that says you can’t predict the markets. Other studies have shown that in long term time frames it doesn’t matter much even if the buying is done at market tops or bottoms, holding quality equities for a long time is all that matters and produces great returns. That has been true for many years in my opinion(depending on your metrics of course), and has a great deal to do with WB-like faith in the system and your own judgements. 50 years ago yes but in today’s market? Not for me, I’ll go with the short term which has an entirely different set of rules.

  • freebird

    Anyone looking for R-sq near 1.0 on this doesn’t understand the concept of risk/reward. The idea is to aim at higher odds for a good return and lower odds for a bad one, not for certainty in outcomes. Here’s an exercise– take all of the dots at PE>20 as one group, and all of the dots at PE<10 as another group. Now run their 10-year subsequent return values as overlaid cumulative distributions. See the difference?

  • pvk22000

    this deserves all caps

    CULLEN, RUN A REGRESSION ON THE SCHILLER 10 YEAR PE TO THE FUTURE RETURNS OVER A 15 OR 20 YEAR PERIOD. THERE IS A R^2 OF 60%. THESE SHORT TERM MEASURES USING ESTIMATED OPERATING EARNINGS OR ONE YEAR OF TRAILING EARNINGS ARE DESIGNED TO SHOW FAVORABLE VALUATIONS. THAT IS EXACTLY WHY SCHILLER USED A TRAILING 10 YEARS.

    I POST THIS COMMENT EVERY TIME YOU SAY THAT VALUATIONS DON’T MATTER OR AREN’T PREDICTIVE. VALUE ALWAYS MATTERS. THAT IS THE WHOLE G.D. POINT OF INVESTING — NOT TRADING PIECES OF PAPER.

  • El Viejo

    What about the overall surprise in the strength of the dollar:
    http://globaleconomicanalysis.blogspot.com/

  • El Viejo

    Long term I care about value, but short term I don’t care. I have no problem taking money from hypsters, be it at a tech IPO or other times. I made money by buying low and selling high on AONE (A123) even though I knew in slower economic times a solar or battery stock would eventually be going down. This is definitely not a buy and hold market and the same strategy applies to most stocks.

  • KB

    Very interesting research! The only question is why for the god sake they used 3Y and 10Y holding period. I believe more prudent approach would use 6m to 10Y with 3 or 6 months intervals. This dynamic would be much more informative.
    I think I saw similar research for S&P (I believe Mish cited it some time ago). It was much more supportive of investing at low P/E…..

  • http://clearlyirrational.com Clearly_Irrational

    You’re basically repeating work that’s already been done. See the Fama & French three factor model for a rigorous study of this issue. They were able to show that over the long term “value” investing does provide a premium that helps explain returns better than just beta alone.

  • El Viejo

    Yeah, you would think that the business cycle would heavily impact this depending on when you start your holding period. In engineering we have something called Nyquist. For example if I am trying to count pulses. I need to measure or examine more than twice as fast as the pulses to elliminate effects of when I start examining.

  • El Viejo

    For valuation buffs (this looked interesting)
    http://www.valuengine.com/pub/main?p=0

  • http://www.pragcap.com Cullen Roche

    I think you’re missing the point. The purpose of investment is to achieve future profits. The purpose of investment is not to buy inexpensive things. If you can’t see the difference there then you probably haven’t done all that well investing over the years. What Wall Street has done is created all sorts of various valuation metrics based on a rear view mirror approach (or their ignorance of future profits) and then packaged this neat little story together where they show you how anyone can become the next Warren Buffett by opening a Wall Street Journal and glancing at the PE ratios of various securities listed. This is a fool’s errand.

    The point is not to say that valuations don’t matter (there is a question mark in the headline). Of course value matters. But the odds of you being able to calculate value relative to future cash flows with the conclusion that this or that company is currently “on sale” in the marketplace is very low. The markets are a hell of a lot more complex than just finding “value”. And in most cases, finding something “on sale” is not even half the battle.

  • http://binaryoptiontutor.wordpress.com/ Options Trader

    I think Nassim Taleb covered this fairly succinctly in his book “The Black Swan“.

    Namely that the vast majority of gains (and more importantly… losses) have fallen essentially on single days / trading events. The vast majority of gains and losses since the Great Depression have taken place on only a handful of days. Be in the right position on one of those days and you do exceptionally well. Be placed incorrectly and essentially never recover.

  • Andyinmiami

    I think John Hussman would vehemently object to the conclusions herein. To say that valuation means nothing is simply ridiculous. P/E, Tobins Q, etc etc; all are invaluable tools in reducing the odds you are buying at the wrong time. C’mon Cullen, you can do better than this.

  • pvk22000

    well, i freely admit to being a retail chump that doesn’t have a lot of investing experience.

    BUT, it is the Schiller CAPE that kept me underweight equities in 2008 and gave me the confidence to begin allocating to equities at the 2009 lows. (It also has caused me to rotate out of equities and miss upside.)

    Reading Rosenberg led me to be overweight on long term bonds and this website gave me the conviction to hold that position through QE2.

    So, I feel like I’ve navigated through the last several years pretty well using a long term approach and my understanding of value. No, I don’t just eat the BS that Wall Street firms put out regarding valuation. I value your thoughts so much that it shocks me when you reference studies like this one as a legitimate analysis of the value approach.

    And what is really an alternative for a retail investor? To try to trade against supercomputers and teams of programmers/mathematicians? No thanks.

  • VII

    I checked out Dracos site. Usual missive stuff pretending to be different but the same. I think this graph graph above has many flaws but I wouldn’t let this get you into a tizzy…these guys seem fairly reasonable and pragmatic. Picking this piece of data above is like a bad soundbit.

    I’m reading alot of posts on Buffet. I’m hearing he has some magical crystal ball no one can come close to or given his wealth would be impossible. Further my GREAT friend Cullen notes, “And honestly, unless you’re Warren Buffett or a hedge fund with access to the Board of Directors or other inside info you’re battling armies of analysts using the same metrics in the same manner. The odds of you finding a diamond in the rough using P/E ratios or similar valuation metrics is frankly, very low.”
    I couldnt’ disagree more with my GREAT friend and other comments here.

    1. Buffet was doing this long before with a paper, pencil, calculator and his tiem. That is it. No computer algos nothing. He was not born wealthy…he was making money looking at the same information everyone else had access to. He did not USE a team of people to become rich. He only had himself. While that is not the case now…..if you look at David Einhorn….it’s the same story. He used fundamental analysis reviewing corporate filings…that’s IT. and it was just him in some dingy shop he was renting. It’s all in his book. Same goes for my favorite Seth Klarman. NONE of these guys were born rich. They got rich by doing exactly what your telling your readers not to do in this post Cullen. They took there time and knowledge and did the work no one wants to do. It’s not sexy, it’s not preatty and it means you have to practice at it. But But But…for god sakes if your going to do anything…wouldn’t you rather spend ALL your time learning this which if you get it will make you Wealthy. Rather than read stupid bull shit graphs that say…WAIT FOR IT….HERE IT COMES…what you pay for an investment DOESN”T MATTER! HUH…What..R U for real? Ok so it’s just the market. Not individual issues. Stocks come in and out of the indicies etc. etc.

    When you listen to Seth Klarman, Buffet, Einhorn talk they all say the same thing….I don’t know why people dont’ learn what we know and focus on things that they don’t believe matter. They have no problem giving away there secrets because they have found that for what ever reason people ignore Grahm Dodd analysis. Saying it’s outdated or the believe wrongly, “you don’t stand a chance against Wall St.” even though the evidence is clear. YOU DO. Crystal Clear unlike this graph above.

  • http://www.pragcap.com Cullen Roche

    See my comment above. I never said value doesn’t matter. I don’t even know how readers are coming to that conclusion when I clearly didn’t state that anywhere in the piece.

  • El Viejo

    My investing history was pathetic until I realized that this is not rocket science and I stopped relying on ‘advisors’. (And I tried most of them) Last year I made 21% on monies invested between Oct10-Oct11. (Mainly due to timely ins and outs into equities – mostly oil and energy. IE: into equities Oct 2010 and out in April and then back in at the dip and back out middle of July ) Since coming here I have improved due to what I believe is a more accurate ‘macro’ picture. In this type of market timing is everything I believe.
    I thought your question mark was pertinent to the perilous times we live in, considering Europe say. I don’t think we are all that bad off in the US given the earnings and I DO look at earnings and PE as part of the picture on individual stocks. I used it when IPSU was coming out of a beaten down condition in the Spring of 2009 (I think). (I’m rattling this off the top of my head.)
    But in this market where people invest based on whether there is a QE3 or not or whether there will be a following to Groupon I just find it more efficient for a person that has to work for a living doing something else to focus on seasonal variations and Biz cycle variations and a good macro picture. I get into this discussion a lot with my brother who leans more to the P & L, earnings and bottom line scenario of individual stocks. He enjoys that type of research and he can have it as far as I’m concerned. It’s just a matter of preference. I love my brother in spite of his predilections.

  • http://www.pragcap.com Cullen Roche

    I am not being a smart ass, but don’t you think it’s a bit of a contradiction to claim that you’re able to time the market using the cyclically adjusted rear view mirror earnings from the last 10 years? You’re essentially claiming to be able to outperform the market using a cyclical indicator in a market timing strategy. I don’t deny that these metrics have value at extremes, but the other 95% of the time you’re just fighting a crowd over the cliff.

    I am not here to claim that value doesn’t matter or that these indicators have zero relevance (I reference them at times). I just don’t buy the notion that valuation indicators are necessarily going to help you achieve alpha most of the time. To me, investing is like being an employer with $100K to burn and walking into a warehouse full of unemployed people. Using a valuation metric like this is the equivalent of going around to the thousands of people and asking them the same questions that 1,000 other employers ALREADY asked them. These people have been poked and prodded using all the same techniques. As an employer looking for value, you have to think outside the box. You have to find something in these people that the other analysts haven’t already discovered. That’s how you find value. Not by checking the same metrics that everyone else is using. And if you can’t find that edge then perhaps you’re better off using an indexing approach. Settle for market returns. Being “average” in this business isn’t half as bad as it’s made out to be.

    In other words, think outside the box. The myth of Warren Buffett has led a lot of people off the cliff.

  • Rharaz

    Thank you Adam!

  • http://alephblog.com David Merkel

    CAPE, q-ratio, and even P/Sales work over long time horizons, like 10 years. Earnings and forward earnings are too volatile to use for any forecasting.

  • http://alephblog.com David Merkel

    To be clear, I agree with you, Cullen.

  • El Viejo

    Yessir! I have missed a little bit of this rally out of concern for Europe, but like they say it’s more important to hang on to what you have. I think a lot of people make the mistake of following a successful strategy and then suddenly some Black Swan pulls the rug out from under their feet. They will focus on what they know or do best or what they prefer or what is easier. To most, positive equity plays are the easiest. I try to watch both sides. I have moved back and forth between bonds and equities. Most don’t realize you can make money in bonds when the 10 yr rates drop from 2.25 down to 1.75. Gary Shilling pointed this out on Bloomberg just the other day. He also mentions it near the end of his book, “The Age of Deleveraging”. And this back and forth pattern has been true for some time now.

  • Juan

    These charts mean nothing, you have to look at individual issues and not at the broad market. See Lakonishok, Shleifer and Vishny (1994)

  • http://www.pragcap.com Cullen Roche

    My, that sounds so easy now doesn’t it? Goldman Sachs has an army of quants looking at individual issues. Do you honestly believe you can find a diamond in the rough using the same valuation metrics that these armies of analysts are using?

  • pvk22000

    i view discipline and patience as my competitive edge. they are formidable advantages and i am tested EVERY SINGLE DAY to maintain them. to echo VII’s comments, this data may be available to everyone, but not everyone is willing to utilize it. you must be willing to wait through periods of overvaluation and to see everyone making money while you are sitting in cash or bonds that pay a pittance. you must also be willing to invest while everyone else is running for the hills. those are not easy things. i believe that if i maintain my discipline, i will modestly outperform the market over long time periods.

    no, i don’t see a contradiction in using CAPE to “time” the market. but i don’t view it as market timing. i view it as waiting for an asset class to be favorably priced.

  • http://www.pragcap.com Cullen Roche

    I don’t see how anyone could use an indicator like this to time the market. For instance, if you were using the CAPE 10 you’d probably agree that a level over 20 is historically consistent with a market that is overvalued. This index has averaged 16.5 over its entire history. I don’t know what you’d call “undervalued”, but if you had started using this index in 1990 you’d have sold equities in 1992 and the market would have remained “overvalued” until some point in early 2009. Anyone using this data has been caught in what looks like an anomaly for almost 20 years! Now, we can blame the Bernanke Put or whatever we want, but pointing fingers and saying that someone moved the goalposts doesn’t matter in making money. You still have to get to where the goalposts are.

    The problem is, you’re trying to compare rear view mirror data with the expectation that today’s investment environment is somehow analogous to the time periods in this data set. Is this data more valuable than using TTM PE’s or even forward PEs? Yes. But that doesn’t mean it will give you a significant edge in formulating an investment strategy. If you have some secret about using the CAPE 10 then I am all ears, but I think it’s foolish to rely primarily on valuation metrics for one’s investment strategy.

  • http://riskandreturn.net Lance Paddock

    As you know Adam, I am a big fan. Great stuff. Hopefully Cullen will read it ;^)

    But, in fairness Cullen seems to skin the cat as well.

  • http://jamesgoodeonthemoney.blogspot.com/ OntheMoney

    Interminable and ultimately un-winnable arguments about what constitutes value drove me into the arms of technical analysis, and I’ve never never looked back.

    To me, a stock is only valuable when its price starts going up. Am I dumb or what?

  • http://www.pragcap.com Cullen Roche

    It is very nice work. But modeling and testing bottoms in the market is very different from being able to put together a strategy based on this info. I personally am skeptical of anyone whose strategy is entirely based on using such metrics.

  • pvk22000

    correct. i would have been heavily underweight equities for most of the 90′s and 2000′s if i had maintained my discipline and strategy (i didn’t start investing until i got out of college in about 2004). my rules would have had me in bonds (corporate and govt). perhaps i would have been in real estate as well — tough to say since that isn’t really in my current rules. (i say current rules because i am certainly still learning and will make subtle changes to my rules as i learn. i am not dogmatic on any of this.)

    so where is the problem?

  • http://www.pragcap.com Cullen Roche

    After adjusting for dividends you missed a 7.7% compound annual growth rate in equities over that period. And that assumes no reinvestment! In other words, this strategy hasn’t just failed over that period. It’s failed with flying colors. Plus, it’s easy to say you would have been in bonds all this time, but the fact is, most value investors have also been saying that bonds are expensive compared to stocks. Using the same sort of phony metrics…..

  • http://riskandreturn.net Lance Paddock

    Cullen,

    I guess it depends on what you mean by entirely. Value gives a lot of information, now what do you do with it? I’ll give some simple examples of how it could. The most basic would be to add more and more bonds to your allocation as stocks become more overvalued. The obvious retort is the one you gave earlier of how this would have led you to sell back in 1992. Well, this doesn’t sell completely, it just adds bonds. Such a strategy does well in the 1990′s and crushes the market since 2000. When stocks get back to near fair value start switching. Over weight as it gets cheaper. Once again, does (and did) very well.

    Of course one doesn’t have to just use stocks versus bonds. It can also be used to over weight areas within the stock market. Thus, using value in 2000 you should have been overweight small and value, REIT’s, bonds, international and emerging (including emerging debt.) You trailed in the late 1990′s, but not terribly, and you clean up in the 2000′s. If you add some long-short based on just those ideas and, well… yeah baby!

    Now let’s add some more aspects. When markets are cheap, emphasize traditional assets and strategies. When expensive, add more emphasis on trading strategies such as Global Macro, Managed Futures, Long-Short and trading strategies in general. That way you only buy equities and credit that is cheap and worth holding. If markets go up anyway the alternative strategies hopefully will do well enough to keep you alive. If not, well they will blow away long only buying of overpriced stocks. This is an example of why the endless arguments between traders and value investors is a bit silly. When value is available, invest more! When it is not, trade more! Can’t we all just get along?

    Add all that together and you have, well, me!

    As for investors who use value and do well with individual securities, Klarman, Greenblatt and many others have done very well. In fact, look at First eagle these last few decades using value and stocks, cash and bonds. No hedges(except occasionally currency and some gold.) They make a mockery of the efficient market guys.

    As for beating the quants, not even that hard. I know people who just by being patient have done very well buying and selling blue chips and some select stocks they favor over the last 12 years using value techniques.

  • http://riskandreturn.net Lance Paddock

    Actually the equity managers have been saying bonds were expensive, and stocks were cheap. They claim to have been looking at value, but they were not really. Nope, real value investors were advising you to overweight bonds most of the last decade. Kicked the markets rear end too.

  • GreedsGood

    The true $64k question is not what the CAPE or even CY2012 PE multiple will be. Rather, it’s the expansion and contraction in the PE multiple that will drive the material market moves.

    Cullen- Have you looked at the recent study from the SF Fed on demographic trends (middle age/old age ratios) and the likely path for PE ratios? I found it very informative. Essentially, their analysis predicts that we’ll see PE ratios continue to drop for the next decade, likely into the 8-10 range.

    When I take a 50k foot view of the market or an individual company’s valuation based on PE, I turn first to the growth rate to understand where the PE multiple is heading (PE/G). In the case of todays global developed equity markets, my assessment is that relying on “normal” PEs is somewhat lazy and inaccurate given the “new normal” in terms of expected future nominal economic growth – which should really be the driver of the PE multiple. Unless one believes that we’ll continue to see 3-4% (real) / 6-7% (nominal) growth rates going forward over the next decade, it’s hard to make the case that the S&P500 is cheap at 13 or 14x.

  • Manolo

    You are implying your edge comes down to psychology, ie patience and discipline.

  • GreedsGood

    From the work of Grantham & Easterling, I believe a case can be made for gauging valuation vs. historical norms (as a “guess) over a horizon of 7 years. That’s an infinite amount of time in today’s volatility, but can be helpful for measuring the entry/exit risk for some investors – like retirees moving into the distribution phase.

    I focus much more heavily on expectations based investing (Malboussin) and sentiment measures to attempt to predict the short-term moves in the markets. The intermediate horizon (1-3 years), seems to be more related to the business cycle. Today is somewhat unique in terms of the macro drivers given the fat tail risks that seem to be growing on the horizon.

  • http://riskandreturn.net Lance Paddock

    Well, I have my opinions about what drives the change in multiples, but actually, I would argue the market is expensive even with normal growth of 3% (4% is not normal.) Current multiples are built on extended profit margins, as Cullen has already noted.

    A value investor would ask first for the market to not be too expensive and then for a discount or “margin of safety.” That way if we don’t get normal growth we get normal returns. If we do get normal growth we get above average returns and so on. Since I believe you are right I demand an even greater discount.

  • pvk22000

    nice strawman

  • GreedsGood

    Just the fact that technical analysis is such a large component of market participation makes it hard to ignore. My model incorporates elements (420 EMA/sell and 280EMA/buy). I also am a big believer in mean reversion and look at the relative performance of market cap and sector indices to understand the market risk appetite and cyclical expectations. I’ll sometimes put on ST trades to catch stocks or indices that are stretched past their 50DMA or have out/under-performed versus their beta-adjusted best fitting index. Seems to hold some value.

    The other side of the technical analysis is really is different fundamental philosophy on investing – momentum based versus valuation-based. I’m more in the latter camp, but often am having to restrain myself before shorting a high-flying momo stock knowing that as long as IBD and price-performance supports the move, it could be months or years before the 50x PE is called into question (despite a slower growth rate.)

  • GreedsGood

    I fully agree, profit margins (and their mean reversion tendencies) are a looming risk for a material drop in the PE. I have run models where I try to normalize EPS based on historical average profit margins and it always looks ugly in terms of PE outlook.

    I’m also a bit perplexed like you as to why the market is perceived cheap at 14x when the GDP is growing nominally in the 5-7% range. I suppose this is due primarily to the fact that corporations have been able to (in recent years) achieve profit growth of 12-15% on average – margin expansion, international & emerging market growth, and currency depreciation have likely contributed to the growth rates well in excess of GDP. But if you go back to the 1920′s, the S&P500 profits are significantly lower which is why during much of history, the PE multiple has also been at much lower levels. **Sorry if I’m not providing exact figures on the historical PE/Growth data. I’ve run the numbers in excel, but have it on my office PC.

  • GreedsGood

    FYI… link to the Fed Reserve study (Aug 2011)

    Boomer Retirement: Headwinds for U.S. Equity Markets?
    http://www.frbsf.org/publications/economics/letter/2011/el2011-26.html

    Conclusion

    Despite theoretical ambiguities, U.S. equity values have been closely related to demographic trends in the past half century. There has been a tight correlation between population dependency ratios, such as the M/O ratio, and the P/E ratio of the U.S. stock market. In the context of the impending retirement of baby boomers over the next two decades, this correlation portends poorly for equity values. Moreover, the demographic changes related to the retirement of the baby boom generation are well known. This suggests that market participants may anticipate that equities will perform poorly in the future, an expectation that can potentially depress current stock prices. In that sense, these demographic shifts may present headwinds today for the stock market’s recovery from the financial crisis.

  • http://riskandreturn.net Lance Paddock

    You are correct on the whole on growth rates. Obviously coming from a trough with record profit margins show huge growth. However, trend growth in earnings has been about 1.5% above inflation. If you take today’s earnings, assume that rate for revenues and then normalize profit margins you get, yuck. A 40-50% decline in equities, That is without a recession, but people rarely accept that analysis because a recession inevitably comes around and gets the blame.

  • http://www.pragcap.com Cullen Roche

    First Eagle’s a good example. Just ran a quick risk adjusted return calc on their value fund. 0.85 Sharpe Ratio. It’s certainly not bad, but in a data set of thousands of value investors we could easily chalk this up as nothing more than an anomaly. And as anomalies go, it’s not even really an outlier. Just an okay risk adjusted return as far as they go…..

  • http://www.pragcap.com Cullen Roche

    A strawman is a misrepresentation of the facts. All I did was point out a fact – you would have missed a 7.7% annualized return. I know everyone wants to jump on the bond bandwagon 17 years too late, but I find it very hard to believe anyone who tries to tell me that they would have moved into a 100% bond allocation following the 1992 sell signal.

    I don’t mean to rain on your parade. I sincerely hope you’ve found a great strategy. And if it works for you then it works. That’s what matters at the end of day. You need to be comfortable with the level of risk you’re taking in achieving the goals you have for your portfolio. I just find it hard to believe that these widely used metrics are helping broad sets of the investment community achieve any real alpha.

  • http://riskandreturn.net Lance Paddock

    The Global Fund has returned 14.23% vs its benchmark that has returned 9.44%. That is since 1979.

    The Overseas fund since 1993 returned 12.06% versus 4.42%.

    Even the US fund has returned 9.38% versus 2.97% since September 2001.

    Three different funds, the only funds they run, over different time frames and massive outperformance. I think that makes it hard to argue that it is an anomaly. I am not a fan of the Sharp ratio, but the absolute returns have been excellent and the drawdowns have been far lower than their benchmarks. Other numbers look much more impressive than the Sharp ratio, but I like the absolute numbers over those time frames as they are not distorted by huge outliers like some managers or funds. They are less volatile to boot.

    However, you are right, we can’t prove it isn’t an anomaly just by the statistics given the time frames. However, then any strategy, including your fund, suffers the same. Thus more needs to be looked at and morem thinking and anlysis than MPT style statistics can tell us. I have been following them since the 1980′s. Their success was predictable. Identifying characteristics ahead of time is quite difficult, they and others fit. Klarman is another example. Finally, I have seen it work for a long time. I don’t think it is an accident that by looking carefully at process I have been able to repeatedly identify managers who over time would outperform using value, even when on the surface they were not that impressive when I put money with them.

    As you often say, getting it right has to count for something. Using value Grantham for example listed almost in order what the best and worst performing assets would be for the ten years from 2000. In fact, his performance at that kind of thing has been impressive over and over for years. The absolute level of performance was extremely close as well. Arnott did something similar. I certainly was looking at similar numbers. Whether people who call themselves “value investors” in the asset management industry used that information well is another thing entirely. However, they got it right, as right as seems humanly possible. So has First Eagle, in different markets, funds and time frames. If we can’t have confidence in that kind of success, then really we should stop talking about anything, since we can’t prove it any better than that.

    Funny, “Value” guys make the same kind of arguments about global macro and CTA guys. In fact they probably have a better argument, since predicting who will do well in Global Macro and CTA’s ahead of time is far more problematic. Generally I am arguing with them about why other strategies are necessary and they start rattling on about sharp ratios, luck and anomalies. When I retort that they can’t prove anything about what they do as well they point out the rationale, rather than proof. Oh well, I guess we all can’t just get along ;^)

  • revelo

    3 and 10 years is much too short for stocks. 20 years is the correct timeframe. And then the correlation of PE10 to returns is very good.

    As for the argument that people who got out of stocks in the 90′s didn’t go into bonds, where exactly do you think they went? There are only 4 true asset classes available to most investors: stocks, bonds, real-estate and cash. Residential real-estate was reasonably priced in 1996. Anyone buying houses or multi-family apartments would have done very well through 2005, at which point they should have sold if they were in a bubble area. People who don’t sell at the top of a bubble are making the same mistake as people who buy at the top of a bubble. Likewise, bonds were reasonably priced in the late 1990′s. Anyone investing in 10 year treasuries from 1996 to October 2008 did very well indeed. Bond holders who didn’t sell bonds in Dec 2008 (or again back in Sep/Oct 2011) are making the same mistake as stock investors who didn’t sell in the late 1990′s or real-estate investors who didn’t sell in 2005-2007. When the market offers a crazy price, take it. The only losers are those who stayed in cash from 1996 to now, but we all know cash is a loser in the long run.

    Right now, stocks are slightly expensive (they were reasonably priced each time the SP500 dipped under 1120), treasuries are very expensive, intermediate-term corporates are slightly expensive, single-family housing is very cheap in many areas of the country, cash is a bad deal. So buy single-family housing if that is an option. Otherwise, buy stocks on the dips and/or corporate bonds.

    If you can’t value asset classes, you shouldn’t call yourself an investor. If you think asset classes can’t be valued by anyone, you are a fool.

  • http://www.pragcap.com Cullen Roche

    Come on Lance. This fund is running a sort of Permanent Portfolio Strategy where they’re globally diversified and have a nice chunk of change in gold which has generated massive returns over the years. For instance, look at this ridiculous 5 year chart of theirs. These are their major holdings. Where is the alpha? It’s ALL in their gold pick with ex-Japan making up the difference. I don’t even have to run the numbers on a fund like this to tell you that “Large Cap Value” is misleading. Back the chart out to 10 years and you just get a super sized view of the emerging market bubble and the gold rally they’ve been a part of. Comparing this fund to a “benchmark” like the MSCI world index or the S&P is just not even remotely accurate. And no, buying gold is never ever a value investment so when it makes up a huge portion of your gains, you can’t really call your fund a “value fund”.

    And just so you can see how incredible this gold bull market has been (and how much it skews a funds returns like this) – if you’d taken 10% of your portfolio at the gold bottom and bought just 10% as these guys did (and left the rest ENTIRELY in cash) you’d have nearly doubled your portfolio. You would have generated almost a 6% annualized return over the last decade with 90% cash. I think if you dig a little deeper into these guys you’ll find that they’re extracting a lot of alpha out of a few positions which aren’t exactly “value” picks.

  • pmarlow

    Adam- Quick note, in the explanatory text for chart #2 (page 7) the colors appear to have been reversed i.e. opposite of what the key on the chart says. You guys have probably already caught this.

    Great report. Thank you!

  • freemarketeer

    I can’t believe how often I hear this criticism of the sell-side. I can only speak for the institutional world, but NO ONE CARES ABOUT ANALYST ESTIMATES.

    http://www.institutionalinvestor.com/Research/3460/What-Investors-Really-Want.html

  • Andrea Malagoli

    I am late to the debate, but my experience agrees with GreedsGood. It is the CHANGE on the margin in PE ratios that matters, not the level. We are in a contracting PE10 cycle, and that is useful information for asset allocators (not for traders) to reduce strategic exposure to equities. So far, the market is playing the cycle by the book. As an asset allocator, one has to work with a 5-7 yrs horizon and ignore the gyrations of the rallies-in-a-bear market, where any upside makes people declare that “this is the beginning of a new bull market”.

    CAPE (PE10) has some use as a strategic indicator. Regular forward or trailing PEs are next to useless.

  • SBG

    They whole reason Shiller created CAPE10 was to smooth the business cycle from the valuation metric. The issue with using forward or TTM data is they don’t account for the business cycle and you get the prettiest metric at the time, lowest valuation at a top and highest at a bottom. So, yes if you use this metric than 20 would be overvalued, however, it is not a timing piece as much as it is an early warning indicator. For me it means I better damn sure pay attention to where we are in the business cycle and plan accordingly. My definition of a secular bear market is a compression in PE ratios. This means to me that the market is still working off extreme valuations and you play to win by not losing. During a secular bear cycle you will have cyclical bulls which will follow the business cycle. If you couple this metric with Lakshman Achuthan’s of ECRI work than you start to build a framework on how to approach the market.