Hi all,
Not sure if anyone has read this book but I thought I'd give a rundown and a few comments anyway to see if I can't rouse a bit of constructive brainstorming.
Basic theory:
Basically, Joel GreenBlatt's strategy runs as follows. You sort two list of the entire market. One by P/E and the other by ROC. You filter them by market cap. This should be big enough so the stocks are sufficiently liquid. He recommends anything over about $50m market cap. You then add the rankings of the two lists and sort by the smallest list. The idea being that you get those business that are the best of both worlds. ie, are quality businesses (high ROC) at reasonable prices (low p/e). As far as portfolio management, you simply buy 2-4 stocks every month until you've done so for a year. If then every month after that you re-balance those stocks that you bought 12months ago (doing some tax tweaking in the process) so you conducting rolling purchases.
In principle I love the theory. I've done a fair amount of reading about simple statistical models outperforming more complex expert driven models. (read some of the blogs on greenbackd.com like this and this).
The idea behind this strategy is that it still works even though everyone already knows about it because people in the market have far too short an investment horizon. Which given the current proliferation of "Wiped of the market"-itis I can certainly identify with.
Secondly, most naysayers will argue that the model is far too simple to be useful. The theory rebuts that because you own between 30-50 quality stocks, the diversification mitigates the outliers and that "on average" you'll be buying quality stocks at low prices.
For example, the ROC could be temporary. ROC is a sign that either competition will compete it away or that there are high barriers to entry protecting it. The strategy contends that on average you'll be ahead.
My questions:
I wonder why for a long term strategy they chose a 1 year holding period. Is this to say that on average, stocks will only take a year to correct their misspricing? Furthermore, is this to say that stocks that turn out to be losses are outliers or are they just perpetually undervalued due to market bias?
Secondly, if a stock is still in your top list after 12 months should you stick it out another 12 months and save some brokerage or cut your losses / take profits there and then?
My backtest:
I decided to see how the strategy went since 1989. I (stupidly) didn't include a dividend adjusted return figure but the results were impressive just the same.
Geometric returns (before dividends) of 18.19% p.a is pretty impressive in my books. I'll get back with the revised returns later on.
The strategy under-performed the market in 7 of those 22 years with only 2 instances of 2year consecutive losses (important because you would have to sit through 2 years of losses and stick with it).
It only physically lost money in 5 of the 22 years with one period of consecutive losses over 2 years.
Again, all this may change once I account for dividends. So I'd disregard most of until its updated.
Not sure if anyone has read this book but I thought I'd give a rundown and a few comments anyway to see if I can't rouse a bit of constructive brainstorming.
Basic theory:
Basically, Joel GreenBlatt's strategy runs as follows. You sort two list of the entire market. One by P/E and the other by ROC. You filter them by market cap. This should be big enough so the stocks are sufficiently liquid. He recommends anything over about $50m market cap. You then add the rankings of the two lists and sort by the smallest list. The idea being that you get those business that are the best of both worlds. ie, are quality businesses (high ROC) at reasonable prices (low p/e). As far as portfolio management, you simply buy 2-4 stocks every month until you've done so for a year. If then every month after that you re-balance those stocks that you bought 12months ago (doing some tax tweaking in the process) so you conducting rolling purchases.
In principle I love the theory. I've done a fair amount of reading about simple statistical models outperforming more complex expert driven models. (read some of the blogs on greenbackd.com like this and this).
The idea behind this strategy is that it still works even though everyone already knows about it because people in the market have far too short an investment horizon. Which given the current proliferation of "Wiped of the market"-itis I can certainly identify with.
Secondly, most naysayers will argue that the model is far too simple to be useful. The theory rebuts that because you own between 30-50 quality stocks, the diversification mitigates the outliers and that "on average" you'll be buying quality stocks at low prices.
For example, the ROC could be temporary. ROC is a sign that either competition will compete it away or that there are high barriers to entry protecting it. The strategy contends that on average you'll be ahead.
My questions:
I wonder why for a long term strategy they chose a 1 year holding period. Is this to say that on average, stocks will only take a year to correct their misspricing? Furthermore, is this to say that stocks that turn out to be losses are outliers or are they just perpetually undervalued due to market bias?
Secondly, if a stock is still in your top list after 12 months should you stick it out another 12 months and save some brokerage or cut your losses / take profits there and then?
My backtest:
I decided to see how the strategy went since 1989. I (stupidly) didn't include a dividend adjusted return figure but the results were impressive just the same.
Geometric returns (before dividends) of 18.19% p.a is pretty impressive in my books. I'll get back with the revised returns later on.
The strategy under-performed the market in 7 of those 22 years with only 2 instances of 2year consecutive losses (important because you would have to sit through 2 years of losses and stick with it).
It only physically lost money in 5 of the 22 years with one period of consecutive losses over 2 years.
Again, all this may change once I account for dividends. So I'd disregard most of until its updated.
