safety

[benhacker]

I'd rather bet on mean reversion or something that's easier and has less of a chance of flaming out or messing up compared to shooting for the moon only to find in 5/10/15 years I fell far short.

 

<<< This. 

 

We must always consider / weight the value of buying 'quality' (long term ROIC >> WACC, with reinvestment, etc... however you define it) with the counter balance that what we may think of as quality may revert to the mean, or worse.

 

To digress a bit, quality and price are a continuum, there is no right answer, but we have to avoid fooling ourselves that we can measure quality very accurately (some can, I agree, but most can't and/or they overrate the "qualitiness" of businesses), price is much easier (although, not always easy).

 

My two cents at least.  Too often investors measure quality on quantitative results over some number of years... sometimes that is right, sometimes wrong.  But when buying, we have to factor in the probability that we will be wrong.

[KCLarkin]

To reiterate, quality statistically outperforms the market especially when combined with value. This has nothing to do with picking the next Microsoft (growth investing).

 

Growth of $1 since 1963 (small cap universe)

Russell 2000 - $269

P/B - $1294

Piotroski (quality and value) - $1462

 

I am not endorsing Cloud's strategy of taking profits as a Margin of Safety but I can certainly endorse buying CSU at 15x cash earnings in 2012 rather than BAC at 0.5x book.

[Tim Eriksen]

To reiterate, quality statistically outperforms the market especially when combined with value. This has nothing to do with picking the next Microsoft (growth investing).

 

Growth of $1 since 1963 (small cap universe)

Russell 2000 - $269

P/B - $1294

Piotroski (quality and value) - $1462

 

I am not endorsing Cloud's strategy of taking profits as a Margin of Safety but I can certainly endorse buying CSU at 15x cash earnings in 2012 rather than BAC at 0.5x book.

 

It seems to me that you should be saying something like "value outperforms and high quality value outperforms that" and not just "quality outperforms" that way you incorporate the essential value component, otherwise the phrase implies that quality alone outperforms which it does not.  I misunderstood what you were meaning.     

[KCLarkin]

It seems to me that you should be saying something like "value outperforms and high quality value outperforms that" and not just "quality outperforms" that way you incorporate the essential value component, otherwise the phrase implies that quality alone outperforms which it does not.  I misunderstood what you were meaning.   

 

The stats I have for long only are: market < quality < P/B < quality value. Some of you are comfortable getting your returns from reversion to the mean. I prefer getting my returns from good businesses and exploiting market psychology to get a decent price.

 

 

 

 

[rb]

The stats I have for long only are: market < quality < P/B < quality value. Some of you are comfortable getting your returns from reversion to the mean. I prefer getting my returns from good businesses and exploiting market psychology to get a decent price.

 

+1

 

Maybe I misunderstand your thinking but is reversion to the mean not the market psychology thinghy? I.e. Market went nuts for a bit and sold the security at a lower multiple and then the multiple reverses to the mean.

 

Also is it just me or does someone else feel that we may be getting punked by cloud?

[KCLarkin]

Maybe I misunderstand your thinking but is reversion to the mean not the market psychology thinghy? I.e. Market went nuts for a bit and sold the security at a lower multiple and then the multiple reverses to the mean.

 

No, I was thinking more of reversion to mean of the underlying business.

Reversion to the mean: BAC has a 6% ROE and it reverts to 12%. Earnings double so P/B doubles. Stock doubles.

Market psychology: Valeant is in the middle of a hostile takeover. There is uncertainty whether they will succeed. Stock sells off 25%. They lose the deal. Stock doubles.

[TwoCitiesCapital]

It seems to me that you should be saying something like "value outperforms and high quality value outperforms that" and not just "quality outperforms" that way you incorporate the essential value component, otherwise the phrase implies that quality alone outperforms which it does not.  I misunderstood what you were meaning.   

 

The stats I have for long only are: market < quality < P/B < quality value. Some of you are comfortable getting your returns from reversion to the mean. I prefer getting my returns from good businesses and exploiting market psychology to get a decent price.

 

Do you mind sharing these stats or the sources of research.

 

Also, another article I found on valuewalk just today further supporting that value > quality.

 

http://www.valuewalk.com/2015/03/rebalance-frequency-value-funds/

 

[KCLarkin]

Do you mind sharing these stats or the sources of research.

 

http://rnm.simon.rochester.edu/research/QDoVI.pdf

 

This one is more fun but answers a different question:

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2553889

 

[TwoCitiesCapital]

Do you mind sharing these stats or the sources of research.

 

http://rnm.simon.rochester.edu/research/QDoVI.pdf

 

This one is more fun but answers a different question:

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2553889

 

Thanks. I've read the 2nd one previously. Will look into this. I read Montier's book last year and I thought that I had taken away that P/B was the best results relative to a ton of value screens and that the unprofitable/loss making companies 1 yr returns generally exceeded the profitable companies arguing against the quality. Maybe I misread/mis-remember it or maybe they're looking at different data? Who knows. I'll re-read the book at some point and let you guys know.

[KCLarkin]

I had taken away that P/B was the best results relative to a ton of value screens

 

Most of the research I have seen is that P/B is not the highest returning value screen. P/B is the default in most academic research so it is most well known. Fama and French believed that all value factors were basically the same so they chose P/B since it was the cleanest.

 

EV/EBIT (or EV/EBITDA) seem to be the current favorite. See Quantitative Value or What Works on Wall Street for comprehensive reviews. I've only read Montier's little book on behavioral investing, so I can't comment on his research.

[oddballstocks]

I had taken away that P/B was the best results relative to a ton of value screens

 

Most of the research I have seen is that P/B is not the highest returning value screen. P/B is the default in most academic research so it is most well known. Fama and French believed that all value factors were basically the same so they chose P/B since it was the cleanest.

 

EV/EBIT (or EV/EBITDA) seem to be the current favorite. See Quantitative Value or What Works on Wall Street for comprehensive reviews. I've only read Montier's little book on behavioral investing, so I can't comment on his research.

 

Let me add a caveat.  As someone who is currently working through backtests I wouldn't entirely trust the data on some of these studies.  I've seen the warts in the historical data.  This extends to the P/B stuff as well.  Backtests aren't as rock solid as many might think they are.

[KCLarkin]

Let me add a caveat.

 

The list of caveats is pretty long. Data quality issues, data mining, re-balancing methodology, tax and transaction costs, arbitrage, liquidity in small caps, long periods of under-performance, confounding variables, limited sample sizes...

 

That's why I prefer a concentrated portfolio of hand-picked fairly-priced quality stocks. My performance might suffer but I sleep well at night.

[cloud]

I simply think that historically numbers and figures disagree with you. It's statistically supported that you can take the stock market universe, divide it into quintiles by P/E or P/B, and each successively lower quintile outperform the higher quintiles and, IIRC, lower volatility. It's also pretty well supported that analysts are categorically overoptimistic in bull markets and too pessimistic in bear markets - or in other words, they're procyclical and spend their careers as momentum chasers.

 

The flaw with this study is that it simply compare price and performance. It ignores the effect of quality.

It's only fair to do it this way: 

If you select a basket of quality stocks and compare it with a basket of cheap stocks, the quality basket will outperform the cheap stock basket by a huge margin.    Of course, the researchers don't know how to select quality, otherwise, they wouldn't do this comparison.  Quality can't be determined by numbers alone.

 

Also as Peter Lynch said: The same PE number means different thing for different business.

During normal economic environment, if the low PE is staying for the long term, it's cheap for a bad reason and this won't give a better performance.

 

Buying quality at any price is not as scary as people think.

Indexers are buying EVERYTHING including many businesses with no earning  at any price. That's real scary.  BUT, a big BUT, they still get 7 to 10% return!

If say let's buy a basket of 30 quality stocks or even 50 stocks with equal weight at any price, isn't it reasonable to expect to make 5 to 10% more than these couch potato?

I always believe with right effort, more work means more reward..

 

 

[cloud]

Of course, both price and quality are important for outstanding investment result. In the real world, these two factors are not fixed. We can choose low price and high quality but the chance of that happening is small.  Therefore, I load up when quality stocks go on sale.  I am a half value investor after all.  :)

 

Currently, I am heavily  loaded with Hibbett Sports, Inc, Michael Kors and Vitamin Shoppe Inc...  10% weight each.

 

 

[UNF2007]

I think this would be an interesting investigation, adding the quality element seems similar to what Greenblat did with the magic formula stocks. But on an individual company basis, I think using quality as your MOS requires a lot of skill, I wouldn't be comfortable doing this, I am more of a pay for the steak and get the sizzle for free person. Reading back through the Greenwald book he has a quality matrix where he compares various returns on capital to cost of capital to come up with franchise multipliers for the EPV. To me it's kind of scary applying a 2x mutiple to EPV based on quality, because if your wrong....

 

I simply think that historically numbers and figures disagree with you. It's statistically supported that you can take the stock market universe, divide it into quintiles by P/E or P/B, and each successively lower quintile outperform the higher quintiles and, IIRC, lower volatility. It's also pretty well supported that analysts are categorically overoptimistic in bull markets and too pessimistic in bear markets - or in other words, they're procyclical and spend their careers as momentum chasers.

 

The flaw with this study is that it simply compare price and performance. It ignores the effect of quality.

It's only fair to do it this way: 

If you select a basket of quality stocks and compare it with a basket of cheap stocks, the quality basket will outperform the cheap stock basket by a huge margin.    Of course, the researchers don't know how to select quality, otherwise, they wouldn't do this comparison.  Quality can't be determined by numbers alone.

 

Also as Peter Lynch said: The same PE number means different thing for different business.

During normal economic environment, if the low PE is staying for the long term, it's cheap for a bad reason and this won't give a better performance.

 

Buying quality at any price is not as scary as people think.

Indexers are buying EVERYTHING including many businesses with no earning  at any price. That's real scary.  BUT, a big BUT, they still get 7 to 10% return!

If say let's buy a basket of 30 quality stocks or even 50 stocks with equal weight at any price, isn't it reasonable to expect to make 5 to 10% more than these couch potato?

I always believe with right effort, more work means more reward..

[cloud]

I think this would be an interesting investigation, adding the quality element seems similar to what Greenblat did with the magic formula stocks. But on an individual company basis, I think using quality as your MOS requires a lot of skill, I wouldn't be comfortable doing this, I am more of a pay for the steak and get the sizzle for free person. Reading back through the Greenwald book he has a quality matrix where he compares various returns on capital to cost of capital to come up with franchise multipliers for the EPV. To me it's kind of scary applying a 2x mutiple to EPV based on quality, because if your wrong....

 

I've read the book on magic formula..  His method should produce satisfactory result giving a large pool of stocks and using equal weight removing the guess work. His screen turns up lots of quality stocks but many of them are mediocre businesses. It's like the hardcore value investing, looking only at the numbers and ignore what the company does, how the company is being run, the future of the market demand etc.

I would cherry pick from them  if I were to build a portfolio. I like companies with little to no debt, stable and growing earning/revenue. This is the starting point to investigate what the company actually does and its market potential.

 

Successful stock picking requires many mental models working together.  That's like becoming closer to a polymath. I followed what Charlie Munger suggested and my investment result and life got a lot better. I am doing what I was not doing 5 years ago and know something I didn't know 5 years ago.

 

 

 

"Mental Models? Latticework? Huh? Do what now?

The phrase ‘Latticework of Mental Models’ comes from Berkshire-Hathaway’s Charles Munger, who is a person that has spent most of his life working out ways to, for lack of a better term, think better.Munger has come to the conclusion that in order to make better decisions in business and in life, you must find and understand the core principles from all disciplines.In short, learn all the big ideas and how they interrelate and better, more rational thinking will naturally follow.This is what he calls Elementary Worldly Wisdom, and using his system of Mental Models can help you succeed in almost any endeavor. "

 

 

 

 

 

 

 

[innerscorecard]

If you select a basket of quality stocks and compare it with a basket of cheap stocks, the quality basket will outperform the cheap stock basket by a huge margin.    Of course, the researchers don't know how to select quality, otherwise, they wouldn't do this comparison.  Quality can't be determined by numbers alone.

 

I thought this was a fantastic point. The difficulty is that there isn't any one factor that's a silver bullet for quality always.

 

The idea of Greenblatt's Magic Formula was to find some rough proxy for quality as a sort of hack to see what would happened if you did a Price and Quality screen. Ironically, Carlisle and Gray's Quantitative Value then found that price alone might just be better than price and quality (not real quality as you said - but a very weak way of trying to find some connected factor to actual quality) in the end.

 

Of course, if you run the Magic Formula or a similar screen, the data shows that investors who cherry pick from these screens or run them non-mechanically underperform those who simply run the screen exactly mechanically due to behavioral bias.

[cloud]

Of course, if you run the Magic Formula or a similar screen, the data shows that investors who cherry pick from these screens or run them non-mechanically underperform those who simply run the screen exactly mechanically due to behavioral bias.

 

It's like most stock pickers under perform the market on average so they discourage any newbies from stock picking. 

The only way to accomplish great result on any new endeavors is first start it and then continuously improve on it.