Announcing Two New Ben Graham Screens

For value investing ideas, you just can’t beat Ben Graham. Benjamin Graham based value screens now make up 5 in my list of predefined value screens.

The two new Ben Graham screens are based on the stock selection criteria that we covered in detail and his famous stock valuation formula from “The Intelligent Investor” which I modified as below.

Ben Graham revised formula

Screen 1: Graham Stock Checklist Screen

Out of the 10 original criteria, I found that the following 4 produced the best results.

  • Criteria 1 out of 10: An earnings-to-price yield at least twice the AAA bond rate
  • Criteria 2 out of 10: P/E ratio less than 40% of the highest P/E ratio the stock had over the past 5 years
  • Criteria 6 out of 10: Total debt less than book value
  • Criteria 7 out of 10: Current ratio great than 2

My thought process of how I came to my conclusions is in more detail in the linked articles.

I then created a test portfolio with the screen results. It’s too early to conclude anything, but so far, the stocks chosen by the screen is considerably outperforming the market.

Screen2: Graham Valuation Screen

I’m excited about this new screen.

You already know that Graham’s method of investing in discounted assets work. But there hasn’t been many discussions proving that the valuation equation actually works.

When tweaked the right way, it certainly does. Here’s the way I interpreted the equation into a screen.

First, the result from the many trials I performed.

Interesting how no stocks would have made the cut in 2001.

Now, how did I create this?

My Thought Process in Creating the Valuation Screen

Narrow Down Companies based on EPS

What I consider to be important when using EPS is to make sure that it is consistent and growing.

You do not want a company with EPS growth of 70%. It will be much too difficult to assess for a screener. Besides, everyone else is screening for huge growth.

The essence of value investing is finding a company at a cheap price. All the companies with huge growth rates are likely overvalued one way or another. What I did was search for, were companies that had temporary depressed EPS growth compared to the mean.

I did this by selecting companies where the 3 year EPS growth was less than the 10 year EPS growth rate. The 5 year and 1o year EPS also had to be positive.

Select EPS Value

Graham used the trailing twelve month (TTM) EPS but I believe I can do better.

There are plenty of sites that offer analyst EPS projections so I used the mean of these EPS estimates. My thought here is that if the current EPS growth rate is depressed but the projected EPS is set to be higher, then surely the passing companies should do better than most.

Specifying the Intrinsic Value Range

I used two versions of the Graham formula

[EPS x (7.5 + 1.5G) x 4.4]/Y and [EPS x (8.5 + 2G) x 4.4]/Y

The formula using 7.5 is considered to be the intrinsic value at the lower end of the range and the 8.5 equation as the upper range value.

Then, simply select all the companies trading with a 33% margin of safety to the lower intrinsic value.

Running the screen, the following list of stocks show up.

While not all the results are accurate, it’s a good start. Definitely lots of potential here.

Don’t forget to view the rest of the value screens.



  • bjorn

    That is GREAT! Did you take down your magic formula though?

  • Thanks bjorn.
    I did take down the magic formula. I need to revise it. Wasn’t working out.

  • Jim

    Jae, I’ve got a theory I’m working on concerning Ben Graham’s Formula. Concisely as I can, I’ve came to the conclusion, I believe thus far, how he arrived at the 8.5 factor. AAA Bonds were returning on average 4.4%. I believe that is a 10 year average. In 1962, from what I’ve investigated thus far, the Dow Jones, which was an indicator he used often in his text books, produced approximately 7.36% over the same 10 year period. Note: from the dates I’m using, I show the Dow averaging approximately 7% but I believe the dates of measurement employed by Mr. Graham were different than the ones I used therefore I’m using an educated guess until I finish collecting my evidence. In any event, 4.4% was the average AAA bond rate and 7.36% was what the market was generating at that time. Collectively, that equates to 11.76%. 1 divided by 8.5 = 11.76% on the nose. The 8.5 factor represent a no growth opportunity during that time and believe that is how he arrived at the 8.5 factor. If you apply this theory to the formula today, we see that in the last 10 years the average of AAA Corporate Bonds have been 5.72% and the S&P 500 has produced an averaged annual return over the last 10 years of 0.42%. Although I wouldn’t use the 0.42% figure as I believe the historical average of 7.5% over the entire history of the S&P is more reliable; part of my research I’m not finished with :) Using these examples, the risk free premium of 0.42% along with the 10 Y average of AAA corporate bonds produce 6.14%. 1/16.3 = 6.135 or 6.14. That would change the equation to V = EPS x 16.3. The S&P 500 is expected to earn $86 per share in 2010 (another measure I don’t put any logical weight to but am using for illustration purposes). Therefore, our equation becomes $86 x 16.3 = $1,401.80. Until I’m finished researching, I don’t see the point in multiplying the AAA bond rate a second time since that factor has already been calculated in the formula to arrive at the risk premium. I believe it unnecessary inflates the end result.

    What I’ve presented is far from being done. Just wanted to share what I’ve been working on as of late and hopefully some of your posters will be able to add some insight into it. All the the best.

  • Jim

    The calculation is very similar to the EPV Model, only the reverse; as well as using Owner Earnings rather than normalized earnings. Quick example:

    Coca Cola: KO

    Current EPS (according to Google): $3.25 p/s * 16.3 no growth = $52.98.

    EPV : Owner Earnings: $5,674,000,000 / 6.14% = $92,410,420,000
    $92,410,420,000 + $13,265,000,000 cash – $26,183,000,000 debt = $131.86 Billion / outstanding shares = $56.84 per share. My EPV was simplified for illustrative purposes. There’s a few more factors that go into it but the “meat” of the calculation has been presented.

    The two valuation techniques represent a very small discrepancy which has to do with the thoroughness of the EPV taking the balance sheet into account. KO is currently trading at $64.27 p/s.

  • Jim

    Did some more work today on my theory. The averaged yield of the Moody’s seasoned AAA Corporate Bonds (30 years) since 1919 – Today is 5.9%. The averaged annual inflation adjusted return of the S&P 500 since 1871 is 6.73%. A no growth PE factor becomes: 1/7.9 = 12.63% (5.9 + 6.73). That changes the equation to:

    V = ((EPS x (7.9 + 2G) x (5.9)) / Today’s AAA Corp. Bond Yield 4.68.

    Example: Coca Cola (KO)

    Averaged Earnings Growth for the last 10 years = 14.10%
    TTM EPS: $3.25

    Formula: V = ((3.25) X (7.9 + 28.19) X (5.9)) / 4.68

    V = $147.88 * 66.67% MOS = $98.59

    KO current market price: $63.62

    Thus far, I believe this would be the correct Graham Formula as applicable today.

  • Jim

    I’ve confirmed that Ben Graham, who first presented this calculation in 1962, must have originally done so between July 1961 and March 1962 when the Long term 30 Y AAA corporate bond yield was 4.4. From 1871 – 1962, the inflation adjusted CAGR of the S&P 500 was 7.40%. Although I believe he used the DJ Industrial Average as his marker, I haven’t determined the DJIA as of yet but the results should be comparable. 7.40 + 4.4 = 11.80. 1/8.5 = 11.76. I’m only off 0.04% which leads me to believe that this minute discrepancy has to do with the exact date he made the measurement. Therefore, it is my belief that it can be logically assumed that this is how he came up with the 8.5 factor in the equation. I also believe to correctly apply it today would be in a similar fashion to how I’ve presented it above. The original equation was:

    Intrinsic Value = EPS * (8.5 + 2G)

    I believe if one brings this up to date, the modified version he incorporated later is not needed because how he arrived at 8.5 is no longer a secret. Therefore, I believe, the new equation should be:

    Intrinsic Value = EPS * (7.9 + 2G)

    The 7.9 factor will change on a month to month basis as the CAGR for the S&P 500 changes as well as the average yield of AAA corporate bonds.

    However, I don’t believe this is a reliable way of valuing stocks. The purpose was only to find how he arrived at 8.5 which I’m satisfied with my discovery.

  • Excellent idea and argument Jim.
    If the 7.36% figure for the Dow Jones return is correct as you cite, then looks like you are spot on with how 8.5 was concluded in his formula.

    And obviously, 7.36% in today’s market is a bit too high so 7-8 looks about right.

    So based on the research that you did, I can also conclude that using 7 in the formula is a good value for the lower range and 8.5 the upper.

  • Jim

    Happy Thanksgiving Jae!

    Here’s my take on this project. I’ve determined that Greenwald’s EPV is an extension to Graham’s original EPV. The Graham formula is comprised of a growth factor (2G), a risk free rate combined with an opportunity cost transformed into a multiple, and the TTM normal earnings of the business.

    After re-reading the history of the stock market chapter of the intelligent investor, I’m fairly comfortable in assuming, thus far, he used the S&P 500 as his opportunity cost and the 30 year Aaa corporate bond yield for his risk free rate. There’s one argument that could be logically made to this but I’ll save that discussion for my website when I’ve finished researching.

    Instead of earnings, Greenwald used buffett’s owner earnings. Greenwald discovered that even though earnings played an important role in deriving market price the equity of the business did too. If you strip away the equity side of Greenwald’s equation, you’re left with:

    EPV = Owner Earnings / (opportunity cost + risk free return)

    Conversely, if we strip away the growth factor of Ben Graham’s formula to match the no growth equation above, the Graham formula would be:

    V = EPS * no growth multiple (opportunity cost + risk free rate)

    The heart of the two equations are identical. If we use $2.50 as an earnings example and an opportunity cost of 15% which translates into a 6.668 multiple, we find the Graham EPV produces a $16.67 share price and the Greenwald EPV produces a $16.67 share price as well.

    Therefore, it makes sense to me, in searching for bargain issues, that one could derive the multiple from a few other sources depending on your comfort level. One could be the standard 15% that you and I both give weight too. Another could be the “no-beta” WAAC, that I personally use. There are some others that could be employed as well but the point is, I believe too much evidence exists that proves the original 8.5 multiple is not a “magic” number. Rather, it makes perfect sense.

    Interesting to me is the genius of Greenwald of how he took the equation to an entirely new level all the while – getting rid of the part of graham’s formula that never made sense (2G) and keeping the formula a truly no growth formula for the investor who doesn’t like to predict, such as me.

    Happy Thanksgiving to all!

  • Jim

    Forgot to include a suggestion. It is my belief, based on these findings, that a comparable discovery could be found basing a screen from Greenwald’s EPV and comparing those results against the Graham Formula. Of course the same factor, say 15%, would have to be used for proper comparison. I don’t believe the in-depth detail of finding maintenance capEX or the proper percent of sales would be necessary for screening purposes. A simple EPV = (Owner Earnings / 15%) + (tangible assets – interst bearing ST & MY debt) / shares out. would be suffice. Just a suggestion.

  • Jim

    * “interst bearing ST & LT debt”

  • Your making some really interesting points. Would love to read about in more detail. Do let me know when you’ve got it up on your site!

    A screen utilizing both the graham formula and a simple Greenwald EPV would certainly yield interesting results.

    Thanks for the idea. I hope to start on it.

  • Hey Jae – The screen you put together is very interesting! I just finished reading The Intelligent Investor and am just going through all my highlights and notes in the book! I noticed you have a copy of the top results loaded on the blog under section title “Screeners.” How often do you update the published screener? Or is it this screener just meant to track the results of the original stocks screened out?

  • @Nate,
    The screens are updated every week.

  • Graham never intended that growth formula to actually be used to evaluate stocks. This is a very common but dangerous misconception.

    See for a scan of the original edition of the concerned page from The Intelligent Investor with a footnote and a warning about this formula.

  • I will have to point out that you have misunderstood Graham. The formula IS intended to value stocks. He just said that it is an approximation which all valuation methods are.