magic formula investing

[frog03]

Updated data on http://www.formulainvesting.com/

 

Considering the amazingly good results over the last 20 years, are any of you using this strategy or something either based on it or similar?  Sure appears to beat essentially all mutual funds and hedge funds...

[Guest kawikaho]

There are several other sites you could use that offers stock screening formulas.  You could check out AAII, or Stingy Investor.  Stingy Investor has a Graham Screen that has returned 467.2% in the past 8 years.  That's roughly 24% CAGR.  And the strategy holds the stock throughout the year, so there's very little turnover. 

 

http://www.ndir.com/SI/articles/1108.shtml

 

I don't know why he doesn't use that strategy in his own client portfolios.  I guess when you hear mechanical investing, people tend to shy away from it.

[Parsad]

I don't like screens at all except for perhaps a starting point for what to look at.  Screens do not take into account many things, such as various other important financial factors that may be precluded from the screen parameters, competitive advantages, economic circumstances, or the value of certain intangible assets.  Screens also exclude other opportunities such as workouts, arbitrage, control positions or fixed income investments.  Cheers!

[Guest kawikaho]

Geez, I guess we're both up late again.  I agree on being wary of blindly using screens, but it makes me wonder when Stingy Investor's own screens far surpasses their client managed portfolios.  I wonder if they ever kick themselves in the butt for not just investing in their screen.

[SharperDingaan]

Re Stingy Investor.

 

'Marketing' returns significantly different from client returns should throw up a bunch of red flags.

It can only happen if the 'market' portfolio is either investing materially differently (holding period, asset mix, securities, fortunate timing, etc.), or the advertised portfolio return itself is false.

 

One lacks 'truth in advertising', the other just lacks truth overall.

 

SD

 

 

[EdWatchesBoxing]

If I'm not mistaken, Norm, the founder of the Stingy Investor, posts here from time to time (at least I remember him from the MSN days). He also writes for a magazine, MoneySense, so I'm pretty sure that he doesn't mean to mislead with the performance information on the Stingy Investor. FWIW, I find that site to be an excellent source of personal finance/investing information.

[Parsad]

No, Norm definitely would not try to mislead in any manner.  He's on here and was on our old board for a long time, and I've met him many times in Toronto, at our dinners, the Fairfax AGM and our own MPIC Funds AGM.  But I think the fact that his screens are difficult to apply to a real portfolio, indicate that screens are a starting point, not a place to finish.  It's the same thing with Greenblatt.  The "magic formula" sounds terrific and sells plenty of books, but you still need the emotional constitution to apply it, and that's why it is exceedingly difficult to replicate.  Cheers!   

[Guest kawikaho]

Especially since some years there's only 1 or 2 stocks the screen filters for.

[NormR]

Happy Summer!

 

I'm on the road at the moment and I'm only connected via a very slow dial up.  So I'll be rather brief.

 

Especially since some years there's only 1 or 2 stocks the screen filters for.

 

That's the big problem with the method.  It's hard to put even a moderately diversified portfolio together based on only a few stocks.  Also, the volatility of a 'few'-stock portfolio will likely be higher than most regular investors can handle.

 

I'll also echo big P's comment that screens really should just be the starting point for further research.

 

Cheers!

 

[Parsad]

Thanks very much for posting Norm!  Cheers!

[netnet]

I would like to put in a dissent here about the screens.  If for example, you take the Stingy-Graham screen or the Greenblatt's screen, you find that you beat 99.9% of money managers. More to the point, who among us has consistently gotten these kinds of returns?

 

We humans have a generally flawed way of looking at things.  If you can find a screen method that gets you Warren Buffett like returns, albeit with way more volatility, then you might do well to get rid of the narrative fallacy and illusions of control and use the screen. Now, if you can't do that at least use the screen as a mark against which to measure your non-mechanical investment choices.  Note however in the situation we were in March of this year, you could clearly find investments that would beat the bogey of the screen(say 20% to 24%).  I would also argue that Fairfax at today's prices could beat the bogey, or at least has a chance of matching it.

 

But the point remains it is damn hard to get to a 20 to 24% per year return, unless your name is Buffett or Watsa!

[Partner24]

If a screen can beat 99.99% of money managers and it's reasonable to think that it will, over the long term, than it's terrific. Furthermore, it could save an investor a lot of time that can be invested elsewhere. So, basicaly I like that formula.

 

The only potential problem that I see with it is the risk of permanent loss. I don't care about volatility that much, but I do really care about not having a permanent loss. Let's say that the stocks it filter get some terrific returns for decades, but one year they get a huge loss for a fundamental reason and you nearly lose all of your capital. Yikes!

 

Cheers!

 

 

[Parsad]

I have yet to meet any individual who has used a screen successfully for years and done extraordinarily well.  It takes more than a quantitative approach.  Cheers!

[Guest kawikaho]

I'd like to see how a screen would perform in various periods.  Alot of these backtests have happened during a period where equities were doing well.  Let's see how they did in the 70's, or how they go forward.  I dunno.  I have tracked Stingy's screen this year, and so far, it's doing fairly ok.  I've noticed one company in the screen, MEOH, also shows up in one of FFH's recent purchases.

[netnet]

I'd like to see how a screen would perform in various periods.  Alot of these backtests have happened during a period where equities were doing well.  Let's see how they did in the 70's, or how they go forward.  I dunno.  I have tracked Stingy's screen this year, and so far, it's doing fairly ok.  I've noticed one company in the screen, MEOH, also shows up in one of FFH's recent purchases.

 

Backtests are fraught with problems, survivor bias and restatement earnings being particularly prominent.  If you believe Greenblatt, who said that he corrected for errors, the key number is outperformance with respect to whatever benchmark you use and the numbers are particularly striking.  The formula really does beat the market substantially.  Most people can not let a computer pick their stocks or tolerate the volatility, when they have such little "control" over their portfolio. 

 

Now to go to Sanjeev's point,

Screens also exclude other opportunities such as workouts, arbitrage, control positions or fixed income investments.

I would say that you certainly can use the a mechanical formula as your base and also do the workouts, arbitrage, etc.  Basically this is what I do these days.  I can not realistically expect to do 20+ per year without "assistance" from mechanical screens.

 

 

[GaliPart]

I have yet to meet any individual who has used a screen successfully for years and done extraordinarily well.  It takes more than a quantitative approach.  Cheers!

 

Parsad,

 

Please don't take this the wrong way, but just because you haven't met them doesn't mean that they don't exist.  In fact, I think that a number of people, let's say Jim Simons as one example, would certainly disagree with your second sentence.

 

          GaliPart

[Guest kawikaho]

Back tests are not fraught with problems, it's the database that is.  If you use a good database, your problems should be minimized.  I think the biggest problem with most databases is survivor bias. 

 

There is another problem with the performance of certain screens that has my internal spidey sense on edge: the time frames might be biased and skew results.  I think that's the biggest one.  Somewhat like Bill Millers record of beating the S&P for 15 straight years. 

 

I'd like to see how a screen would perform in various periods.  Alot of these backtests have happened during a period where equities were doing well.  Let's see how they did in the 70's, or how they go forward.  I dunno.  I have tracked Stingy's screen this year, and so far, it's doing fairly ok.  I've noticed one company in the screen, MEOH, also shows up in one of FFH's recent purchases.

 

Backtests are fraught with problems, survivor bias and restatement earnings being particularly prominent.  If you believe Greenblatt, who said that he corrected for errors, the key number is outperformance with respect to whatever benchmark you use and the numbers are particularly striking.  The formula really does beat the market substantially.  Most people can not let a computer pick their stocks or tolerate the volatility, when they have such little "control" over their portfolio. 

 

Now to go to Sanjeev's point,

Screens also exclude other opportunities such as workouts, arbitrage, control positions or fixed income investments.

I would say that you certainly can use the a mechanical formula as your base and also do the workouts, arbitrage, etc.  Basically this is what I do these days.  I can not realistically expect to do 20+ per year without "assistance" from mechanical screens.

 

 

[manualofideas]

Here are some magic formula screen results (based on forward earnings estimates):

http://www.manualofideas.com/blog/2009/08/10x45_bargain_hunter_screen_re.html

[Parsad]

Parsad,

 

Please don't take this the wrong way, but just because you haven't met them doesn't mean that they don't exist.  In fact, I think that a number of people, let's say Jim Simons as one example, would certainly disagree with your second sentence.

 

You could very well be correct.  I'm still not convinced of Simon's results, so you'll have to view me as a heretic.  Cheers!