Grading the Gurus

I used to think of myself as contrarian, but after completing my most recent research project, “Grading the Gurus” (which I presented at the last MBIIM) , I now realize that I only appear to be a contrarian. I’ll spare you the discourse on how I came to that for now, but I’ll make sure to include it in my closing remarks.

For those who were unable to join us last Sunday, I thought I would summarize the presentation and the following discussion. Without further ado…

Like my previous post says, “I have often wondered if it makes any sense to pay attention to investing gurus.” And there certainly are a lot of them. Most of which seem to promise you that they’ve found the “secret” to easy money, whether that be a method of valuing companies or assessing the market’s future direction. However, evidence suggests otherwise as it has been proven that 85% of mutual funds have underperformed “dumb” index funds over the last 40 years. This means that all those fancy folks that went to fancy schools and wear fancy neckties are not as smart as “passive” investors. Therein lies the problem…

In heeding the investment advice and methods developed by others, the problem is that most people do (or advise others to do) what woulda worked in the past, while very few can reliably do otherwise in a forward-looking fashion.

This phenomenon persists for any number of reasons including fees, lower portfolio volatility demonstrated by mutual funds (i.e., less risk = less reward), and market impact (i.e., traders trading ahead of the funds). While these limiting factors are particular to money-managers, the adage “past performance is not indicative of future performance” also applies to guru investing, because (a) trading costs are always present; (b) replicating a genius’ brain is problematic; and, (c) markets tend towards efficiency.

Why does any of this matter? I think this matters because it behooves me to have realistic expectations regarding expected returns if I were to adopt some lessons of the legends. For example, Ken Fisher makes an intriguing argument that price/sales ratios are more meaningful metrics of value than price/earnings. I want to know if any data supports this (besides the historical data that each of them had access to at that time). I want to know which, of any, of these gurus have “out-of-sample” predictive power. I also want to be able to test the gurus not on their actual returns, but rather on their methods since that is a better test of their skill (as opposed to genius). And finally, I want to reduce my decision making complexity by placing as many of the experts as possible, if any, into one or more of the following categories:

  1. savant: someone who clearly has the brains to pick stocks well, but hasn’t really defined their methods in any replicable way.
  2. lucky bloke: someone who made achieved great returns, but not necessarily because they were smarter or more skilled. Probability theory predicts that their will be plenty of these individuals floating around.
  3. intellectual fraud: I think there is a lot of this going around. Although most pundits dole out bullshit in manageable quantities, some of them, especially the “self proclaimed” gurus, spout it like gospel.
  4. curve-fitter: This applies mainly to the academic crowd, whereby a researcher reveals a seemingly promising trend or anomalies, but that ultimately fails to deliver in forward looking fashion.

To re-iterate, I am looking to differentiate skill, as measured by the efficacy of the gurus’ methods in an out-of-sample data-set (i.e., performance estimates that were created using methods that existed predating the underlying data). I measure this not only by notional return but also by robustness in performance. I think this is important because skill can be taught and learned by people like you and me; genius may be more problematic.

PiotroskiDossier OShaughnessyDossier ValideaDossier ZweigDossier LynchDossier NeffDossier GreenblattDossier FisherDossier GrahamDossier BuffetDossier DremanDossier FoolDossier

See the below PDF document for dossiers on the 12 guru evaluated:

All analysis on guru performance was conducted using publicly available data from The benefit to using this is that my results are 100% transparent; anyone can replicate my result. The limitations are the method of computing back-tested performance is not 100% transparent, and that parameters of the historical performance are somewhat rigid. All things taken into consideration, I believe that it is nonetheless enlightening to take a fresh perspective on forward-looking, publicly available data that has been tracking the “gurus” for over 10 years.

‘Quants’ use a multitude of performance metrics to weigh performance. I think that 2 matter the most: total returns; and, Jensen’s alpha.

  1. Total Return: Most people know how to interpret percentages. I am taking compounding into account, and thus use Compounded Annual Growth Rate (CAGR) to measure total return.
  2. Jensen’s Alpha: alpha in my field translates to risk-adjusted return in excess of some benchmark; ‘predictive power’ in other words. It has its roots in the Capital Asset Pricing Model (CAPM), which is perhaps the most elegant and most misused mathematical formula in all of finance. Jensen’s alpha is defined by the following formula:

    in which α is risk-adjusted return; r(a) is the return on an asset/portfolio; r(f) is the risk-free or discount rate, β is the OLS coefficient between r(a) and its benchmark r(m); and, r(m) is the market’s return.

In ranking the gurus, I am ultimately looking for both the highest returns and the greatest degree of robustness. I am roughly defining robustness as model’s ability to withstand a change in one of the parameters with little to no degradation in the result. Improved results as a consequence of changing parameters is preferable to degradation, but consistency is the key. I have to confidence that a historical model is not fine-tuned and can withstand me changing the number of stocks I would have held or the length of time I would have held them.


Matlab Gurus Legend

Total Return:



Annualized All-Time Return
Rebalancing = Monthly Rebalancing = Quarterly Rebalancing = Yearly
Model Portfolio Based on Inception Date # Stocks = 10 # Stocks = 20 # Stocks = 10 # Stocks = 20 # Stocks = 10 # Stocks = 20
Growth Investor Martin Zweig 7/15/2003 10.60% 12.10% 12.40% 12.90% 8.80% 7.40%
P/E/Growth Investor Peter Lynch 7/15/2003 9.80% 15.30% 10.30% 13.40% 7.90% 6.60%
Small-Cap Growth Investor Motley Fool 7/15/2003 15.90% 11.20% 8.90% 8.50% 8.30% 4.80%
Earnings Yield Investor Joel Greenblatt 12/2/2005 9.00% 7.20% 10.10% 6.30% 1.30% 2.80%
Momentum Investor Validea 7/15/2003 7.70% 8.90% 9.70% 11.50% 11.60% 9.40%
Value Investor Benjamin Graham 7/15/2003 15.50% 17.60% 14.90% 14.00% 15.00% 14.10%
Price/Sales Investor Kenneth Fisher 7/15/2003 13.50% 12.20% 14.50% 10.30% 11.80% 10.40%
Patient Investor Warren Buffett 12/5/2003 7.20% 5.30% 4.60% 5.20% 4.00% 2.50%
Low PE Investor John Neff 1/2/2004 3.20% 5.30% 3.50% 7.10% 0.00% 0.90%
Contrarian Investor David Dreman 7/15/2003 7.00% 7.00% 5.00% 5.50% 6.40% 3.40%
Growth/Value Investor James P. O’Shaughnessy 7/15/2003 9.90% 10.70% 9.50% 10.90% 7.30% 8.00%
Book/Market Investor Joseph Piotroski 2/27/2004 4.40% 6.10% 8.40% 7.40% 9.50% 12.00%


Annualized All-Time Alpha
Rebalancing = Monthly Rebalancing = Quarterly Rebalancing = Yearly
Model Portfolio Based on Inception Date # Stocks = 10 # Stocks = 20 # Stocks = 10 # Stocks = 20 # Stocks = 10 # Stocks = 20
Growth Investor Martin Zweig 7/15/2003 4.88% 6.11% 6.62% 6.96% 3.02% 1.62%
P/E/Growth Investor Peter Lynch 7/15/2003 3.54% 8.98% 4.14% 7.30% 2.12% 0.55%
Small-Cap Growth Investor Motley Fool 7/15/2003 9.85% 5.26% 2.85% 2.51% 2.52% -1.09%
Earnings Yield Investor Joel Greenblatt 12/2/2005 4.98% 2.99% 6.01% 2.09% -2.60% -1.18%
Momentum Investor Validea 7/15/2003 2.35% 3.82% 4.46% 6.32% 6.15% 4.11%
Value Investor Benjamin Graham 7/15/2003 9.13% 11.23% 8.69% 7.68% 8.90% 8.11%
Price/Sales Investor Kenneth Fisher 7/15/2003 7.13% 5.88% 8.34% 4.09% 5.86% 4.35%
Patient Investor Warren Buffett 12/5/2003 2.06% 0.40% -0.55% 0.30% -0.95% -2.40%
Low PE Investor John Neff 1/2/2004 -1.98% 0.04% -1.63% 1.84% -5.04% -4.41%
Contrarian Investor David Dreman 7/15/2003 0.41% 0.47% -1.80% -1.03% 0.73% -3.03%
Growth/Value Investor James P. O’Shaughnessy 7/15/2003 4.55% 5.41% 3.88% 5.61% 1.58% 2.38%
Book/Market Investor Joseph Piotroski 2/27/2004 -1.27% 0.22% 2.69% 1.48% 4.84% 7.42%

… and the winner is, in my book, good ol’ fashioned value-investor Benjamin Graham. His relatively simple method of picking stocks with the highest ‘intrinsic value’ relative to their current market price clearly outperforms all other gurus for the 10 years of data. Although the Motley Fool method outperforms in terms of total return and alpha when the parameters are set to 10 stocks held monthly, it shows a steady decline in performance the greater the number of stocks held and the longer the re-balancing period. Clearly, it does not meet my ‘robustness’ criteria.

Not only does the Benjamin Graham method out-perform all-time, but it also shows a fairly consistent rate-of-return across the 10 annual periods.

The Benjamin Graham Method
Because Graham wins by a brutal TKO, I thought it would be only fair to go into detail regarding how his method works. As you read the below criteria, it might be helpful to ask yourself throughout, “what does this parameter really mean?”, and “what is he trying to get at?”

The method:

  1. Exclude tech or financial companies (not to imply they are bad investments, but rather Graham realized that rapidly growth companies need to be systematically valued differently)
  2. Adequate size: Trailing twelve month (TTM) sales volume >= $340 M
  3. Shareholder’s residual claim on assets: Current Ratio >= 2
  4. Long-term solvency: Net Current Assets >= Long-Term Debt
  5. Consistent profitability: 30% increase in Earnings-per-share (EPS) over last 10 years; no negative EPS in the last 5 years
  6. Meets a reasonable value threshold: P/E * P/B <= 22
  7. Rank by Relative Value:
    RV = IV / Price
    IV = EPS * (8.5 + 2g) * 4.4 / Y
    where RV is the relative value; IV is the intrinsic value; g is the 5-year EPS growth rate; Y is the current yield on AAA corporate 10-year bonds

I took on this research project because I really don’t fully trust just anyone’s word on what to believe. It’s not like I automatically take the opposite side of any argument just to be a philosophically consistent contrarian, and it’s not even that I don’t care. I just like to come to my own conclusions when presented with facts (which may or may not coincide with institutional beliefs). Although that sounds an awful lot like me not caring, I truly believe it is important for ‘enterprising investors’ to evaluate the possibilities because there is so much crap out there. Who knows what to believe or who to trust? Other than that, yeah, I guess I really don’t care.

Anyway, hopefully this post will allow you to evaluate some of the world’s foremost investing gurus and come to your own conclusions as well. I recommend that individuals judge the gurus not solely on who got the best returns in notional or in real-life, but rather about who’s got the philosophy which is most consistent with one’s own style and concept of ‘reality’. Maybe it’s a blend of multiple styles, and maybe you think that they are all bologna. It’s your call.