Message Font: Serif | Sans-Serif

No. of Recommendations: 1
In this one minute video Charlie Munger enumerates BRK's four-point business evaluation formula:

To paraphrase his remarks, the business must:
• Be understandable to the investor
• Have a durable competitive advantage
• Be managed by a team possessed of integrity and talent
• Be attractively priced

Regarding the final point, here's another brief video that claims to
represent Buffett's valuation criteria:

The formula presented is Value / Price
where Value = Earnings + Dividends
and Price = Price/Book

According to the video, when this formula evaluates to...
< 2 the price is likely to decline
> 10 the price is likely to rise
between 2 and 10 is average

Given a watchlist that seeks to satisfy Charlie's first three criteria, what do others think of this formula, and what other evaluation measures might apply.

In addition to advocating high ROE, Mungo has previously suggested that Debt should ideally be less than 5 times Annual Earnings.

Your input in this regard would be much appreciated.

Tom
No. of Recommendations: 1
The formula presented is Value / Price
where Value = Earnings + Dividends
and Price = Price/Book

According to the video, when this formula evaluates to...
< 2 the price is likely to decline
> 10 the price is likely to rise
between 2 and 10 is average

Given a watchlist that seeks to satisfy Charlie's first three criteria, what do others think of this formula, ...

It is a nonsensical formula, if I am understanding it correctly.

First, it makes no sense to add earnings to dividends, since dividends are a part of earnings. Second, if Price = Price/Book (ignoring the inelegant re-use of the same word to mean 2 different things), then it is a ratio (for instance, Berkshire has a ratio of about 1.35), whereas the numerator of the formula, Value, is the absolute amount of earnings + dividends. How are we to apply this to Berkshire, for instance? Earnings are about \$9 billion per A share, or \$6 million per B share, so Value/Price is either 7 billion or 4 million, depending on which type of share we're talking about.

Regards, DTM
No. of Recommendations: 0
Earnings are about \$9 billion per A share, or \$6 million per B share, so Value/Price is either 7 billion or 4 million, depending on which type of share we're talking about.

I mean earnings are \$9,000 per A share (not \$9 billion, unfortunately), and \$6 per B share, perhaps obviously. Doesn't change my point, but ...

dtm
No. of Recommendations: 1
I interpreted it as (earnings yield + dividend yield) / (P/B)
Double counting dividends makes sense since "a bird in hand is worth two in the bush".
Limits are a bit too much. For example, for a P/B of 2 (modest, e.g. a value stock), EY + DY will have to be > 20%, which is a lot.
Also, B in the P/B can go down thanks to sustained buybacks (and DY will be understated) so you need to make some modifications in the presence of significant buybacks.
All in all, about as good as any other formula.
No. of Recommendations: 0
It is a nonsensical formula, if I am understanding it correctly.

My apologies for poorly expressing the formula. As I understand it, all terms in the equation are per share, and Earnings refers to retained earnings.

Tom
No. of Recommendations: 0
Limits are a bit too much. For example, for a P/B of 2 (modest, e.g. a value stock), EY + DY will have to be > 20%, which is a lot.

What would your thresholds be for reasonable P/B and EY?

All in all, about as good as any other formula.

I suppose any formula is "good" as long as it's mathematically sound. What other investing formulas do you feel are worth considering?

Tom
No. of Recommendations: 28
I suppose any formula is "good" as long as it's mathematically sound.
What other investing formulas do you feel are worth considering

Whether it makes economic or mathematical sense is debatable, but I can certainly
recommend the following formula because it makes money year after year:
- sort them by product of dividend yield and earnings yield.
This implicitly skips all stocks not paying dividends.
- keep the best 20-50 stocks on that sort. Best 35 is traditional but it doesn't matter much.
- of those, buy the few stocks that have risen most in price in the last year or so (12-16 months)
It's best to ignore any price change in the last 2-6 weeks or so, but not critical.
Anywhere 3-10 stocks is good, but 4-5 is the usual choice.

Hold those stocks for 3-13 weeks and repeat. A monthly check is traditional but again not critical.
I've been using every 2 months lately. The picks don't actually change that often.

This would have returned around 20%/year since 1986, including the
last decade which is entirely after the method was popularized.

Jim
No. of Recommendations: 0
If this method has been popularized, is there a name for it? Also, any analysis of why it works?

Thanks.

-Rubic
No. of Recommendations: 0
I hope Jim is kidding. I have not been reading this board enough to know if we have truly entered the heart of darkness or just are having a good laugh at uneducated spammers. Original formula is complete BS. Any first time readers should be careful of anything that starts with something to the effect "Warren Buffett uses this formula...." I fault my self for clicking on the you tube video.
No. of Recommendations: 0
Ah, I jumped into this thread not realizing April Fools comes a couple days early in Monaco.

-Rubic
No. of Recommendations: 24
If this method has been popularized, is there a name for it? Also, any analysis of why it works?

Yes, there is a name for it.
Its designation over at the Mechanical Investing board is "YLDEARNYEAR"
since it looks at dividend yield, earnings, and one year performance.

No, it's not an April fools' joke.
Yes, the returns really have been market beating for quite a long time.

To get a feel for it, go to this link, click "run", and see the results.
http://www.backtest.org/ST%28YLDEARNYEAR%2915
This gives the year by year returns and shows the actual stock picks
for one particular possible set of trading dates since 1986.
The results 1986-2002 are with hindsight, used to test the original model, so they should be taken with a big grain of salt.
In this era the screen beat the S&P by 16%/year before trading costs (28% to 12%).
The results from early/mid 2003 to date are since the screen was published,
so that represents clean out-of-sample validation of the model.
The results show it beating the S&P by about 17%/year in this era (24% to 7%),
which lends credence to the notion that the original backtest wasn't a fluke.
It has beat the broad market in a given 12-month period about 74% of the time.
Based on the backtest, which seems to conform well to reality, that's about a 1/4
chance of underperformance averaging -10% and a 3/4 chance of outperformance averaging +28%.

The screen was originally inspired by a value investing article at a site
called the Stingy Investor, which noted that the "Value Ratio", being
dividend yield divided by P/E ratio, was a not-bad way to find good value.
(this addresses only the earnings and dividend portions of the screen, not the price momentum final step)
A geeky side comment: the P/E ratio is not numerically well defined,
having problems with low, zero, and negative earnings, so it's always
better to talk about its inverse, earnings yields instead. Thus the Value Ratio
goes from divyield/pe to divyield*earningsyield = (div/price) * (eps/price).
This is the original article, conveniently still up
http://www.ndir.com/SI/articles/1099.shtml
Most of that article was itself first published in October 1999.
This is the thread in which the screen was formulated and refined in June 2003
http://boards.fool.com/interesting-metric-19157577.aspx?sort...
I believe this is the first post to formulate it pretty much as it is now http://boards.fool.com/monthly-current-yield-1-current-pe-ra...
(it was slightly simplified, no longer requiring that the stock have
a current Timeliness ranking from Value Line, which is in effect just
a check to eliminate firms undergoing mergers, very new firms, etc.
It turns out to work better if you just leave those in the mix and not worry)

The number of stocks taken at the first step of the sort at 35
wasn't tuned at all...it was simply taken from the original article
which looked at the TSE35 index. It turns out to be a good choice for US stocks.
(best 35 by value ratio, then best handful of those 35 by momentum)

I haven't run this screen continuously, but I have averaged about 20%/year
in real life with it during the periods that I have used it.
Better than a kick in the head.

Each week the current picks are posted over at the MI board for anyone to use.
e.g., see this post for last week http://boards.fool.com/rankings-25mar2013-30605827.aspx
If you scan down till you see YLDEARNYEAR, you'll see the current top picks are AB KKR AINV STX UNTD.

I first mentioned this screen on this board March 20, 2009 in
this post http://boards.fool.com/maybe-if-they-are-both-feeling-partic...
In the 4.0 years since that date, allowing 0.4% trading costs it has returned 36.8%/year.
Of course we have had a big rally in that period...the universe of
stocks from which it was selecting also returned 32%/year.
Still, it's impressive that a heavily value oriented approach still
managed to beat the market meaningfully during a strong rally.
(how well did the hedge funds do that are in the wager with Mr Buffett?)
If you beat the market by 5%/year for a long time, you tend to get rich at some point.

I'm not trying to convert anyone to the quant heresy, just passing on the information.
Based on the data, it seems to work pretty well.
Particularly suited to tax-exempt accounts as there is significant
turnover so those who pay capital gains tax will pay it regularly.

Jim
No. of Recommendations: 0
I'm not trying to convert anyone to the quant heresy, just passing on the information.
Based on the data, it seems to work pretty well.

Jim, I'm wondering if you have any links (or personal experience) with someone who 1) identified a good back-tested formula, 2) publicly began an investment using the approach, and 3) had a long run of success with the formula.

I have seen so many great formulas--some with very impressive back-dating--yet I can never quite make the leap of faith.
No. of Recommendations: 2
Jim, I'm wondering if you have any links (or personal experience) with
someone who 1) identified a good back-tested formula, 2) publicly began an
investment using the approach, and 3) had a long run of success with the formula.

I can't speak about the wider population of world investors, but there are certainly such people over at the MI board.
e.g. this post
http://boards.fool.com/doesnt-work-is-reflected-in-the-poll-...
Outperformance of 5.91%/year for 13 years, not bad.

As with Elan's results, I've had moderate but useful outperformance on average over time.
A few percentage points a year of tailwind makes a fair bit of difference over time.
It has worked a lot better since i added a wee bit of timing.
If you can take some usefully large chunk of the pain out of most bear
markets by sitting on cash (not hard), the long run average results get a lot better.

Jim
No. of Recommendations: 2
I can't speak about the wider population of world investors, but there are certainly such people over at the MI board.
e.g. this post http://boards.fool.com/doesnt-work-is-reflected-in-the-poll-...

FWIW Another one from today.
http://boards.fool.com/since-jan-2006-my-mi-cagr-as-of-33120...
7 years beating the S&P by 8.1%/year. Pretty good.

The hard part is sticking with it, and the hardest part of that isn't
that the bad times are any worse than with any other investment strategy,
but rather that they are at different times.
Losing 15% occasionally is not hard when everybody is doing the same.
It's hard when everybody else is making money.

Jim