This post has been about 2 weeks in the making. A couple of weeks ago, we were asked in a comment if we could elaborate a little on our valuation process. We had been using a complex mathematical formula and system to determine value that took into account a number of different approaches. As the system was difficult to explain, we always had a stance that we would not discuss the details of how we came about our intrinsic value numbers. However, when I was thinking about how to respond to the latest question, I also began to wonder if we could find something that was easier to explain.

I was led back to Graham’s The Intelligent Investor, where I found a formula that I had originally passed over as it is in a section that refers to “Growth” stocks. The first couple of times I read the book I thought of the formula as a way to look at high growth stocks, but this time, I saw it differently. If it worked for Graham with growth stocks, why couldn’t it be tweaked in order to work for all companies? I then set to work on “modernizing” Graham’s original formula.

Here is the text from The Intelligent Investor that includes the formula (found on page 295 of 2006 updated edition – chapter 11):

“Most of the writing of security analysts on formal appraisals relates to the valuation of growth stocks. Our study of the various methods has led us to suggest a foreshortened and quite simple formula for the valuation of growth stocks, which is intended to produce figures fairly close to those resulting from the more refined mathematical calculations. Our formula is:

Value = Current (Normal) Earnings X (8.5 plus twice the expected annual growth rate)

The growth figure should be that expected over the next seven to ten years.” (A footnote then explains that the formula does not give the “true value” but approximates)

So Graham provides us with a base formula to approximate value. The next step was to find the correct figures to input into the formula. You’ll notice that Graham has specified that we should use “normal” earnings in this calculation. This is one of the most important things that Graham teaches – do not base your judgment on current period earnings that could be abnormally inflated or deflated. Instead, base your judgment on the earnings that the company has normally achieved and can be expected to achieve in the future.

But how do you get normal earnings in a simple manner? One of the things we at ModernGraham have done since the beginning is use a weighted average of the last 5 years’ earnings per share. We use a weighted average because we do feel that some bearing must be placed on the current period even if it is an abnormal result. Accordingly, our average increases in weight each year. We refer to this weighted average earnings per share as the EPSmg.

From there, the next question is where do you get a growth rate for the next seven to ten years? You can do that a number of ways. One, you can find an analyst forecast and plug their rate in. We don’t like to trust analysts because they are almost always wrong and they can hardly ever agree. Instead, we calculate our own expected growth rate using a weighted average on the previous 5 years again. Then we throw in one of our safety screens by multiplying the average by 0.75 (effectively decreasing it by approximately 25%). Sometimes this still gives us an outrageous figure. As a result, we limit our growth rate to a level between -4% and 15% per year. We feel that any company that is expected to shrink by more than 4% per year is unsuitable for investment anyway, and no company should be expected to achieve higher than 15% growth for seven to ten years. That’s not to say higher than 15% growth may be achieved – it can and it has been by a lot of high growth companies – but we do not want to put ourselves at risk of it not being done. If we put the cap at 15% and still find the company to be undervalued, then our results will be even greater when a higher rate is actually achieved.

You’ll notice that there are a couple of implications that result from our cap of the growth rate. One, it means that we feel a zero-growth company would be worth at least 8.5 times normal earnings. This is the same as what Graham felt and could be a result of an approximation of the present value of a perpetuity of the normal earnings. Two, the cap means that we do not feel comfortable paying more than 38.5 times normal earnings. In fact, we do not feel comfortable for the defensive investor at more than 20 times normal earnings.

After “modernizing” Graham’s original formula, I then set about testing the results of following it. In order to do this easily, I made a couple of assumptions:

- I looked only at the current 30 components of the Dow Jones Industrial Average.
- I did not consider whether the companies passed the defensive and enterprising investor guidelines we require.

With the assumptions made, I came up with a valuation for each company on a quarterly basis going back to the first quarter of 1995. I then looked at the company each quarter to compare the price to the value. If the price was below 75% of the value determined by the formula, I considered the company a “buy.” If the price was above 110% of the value, I considered the company a “sell.” Next, I created a mock portfolio of $100,000.00 and set about following the recommendations of the formula. When a rating changed to “buy,” I “bought” and when it changed to “sell,” I “sold.” I set a target weighting of 5% per company and left the remainder in cold hard cash when I did not achieve a full level of investment. Some very interesting things happened.

First, the portfolio held strong at a level of between 70 and 90% invested until Q4, 1998. At that point, the cash level increased to 46% and stayed between 40 and 60% until the end of 2003. After that, the cash level slowly decreased until 2006 when the portfolio reached a full investment level. As a result, the portfolio missed out on the over inflated market of the very late 90s and the subsequent recession and drop in value. Then once the price dropped relative to the value again, the portfolio increased the investment levels to take advantage of the low prices.

Second, the portfolio as a whole outperformed the index it was up against, the DJIA. On an annual basis, the Dow averaged an 11.5% gain over the period, compared to the portfolio’s 12.9%. In addition, the standard deviation of the Dow was 17.1% compared to 11.1%. So the portfolio achieved a higher rate of return with a lower level of risk.

The overall result was that if you had invested $100,000 in the Dow and $100,000 in the portfolio I tested, you would have $323,000 through the Dow and $408,000 through the portfolio today. I’d say that’s a significant difference, and here at ModernGraham we will be commencing use of the “modernized” formula that Graham provided in The Intelligent Investor for the foreseeable future.

Hi Ben, I have been working on something similar so was very interested to find your post!

What tool do you use to do the backtesting of your formula please?

Thank you very much and best of luck with your research!

Adam

One question: Whats the basis for choosing the ’8.5′ figure in your value calculation formula above ?

Is it related to the long term bond yield rates in some measure ?

Also, isn’t this value discovery formula just another way of getting at the PEG ratio of a company ?

Thanks

Sudeep

How do you actually weight earnings and growth?

While I’m on the subject, I was wondering if you could/would provide a short glossary of terms that you use in your “Defensive”/”Enterprising” questions (e.g., “current ratio”).

Of course, it wouldn’t hurt if I just went back and reread The Intelligent Investor a few more times.

Many thanks … love the site…

This is very interesting and helpful.

I have a few questions:

1. Did you record the number of “Buys”, “Sells”, etc. which occured during the backtesting? If you had that quantity, it could easily be mulitplied by a certain transaction price to come up with a net profit.

2. Are you using a spreadsheet for this analysis? If so, are you aware of the totally great tool called the smf_addin at Yahoo Groups? It provides functions that create new excel “commands” that make it easy to pull all kinds of data from several different databases!

3. It seems that the site’s authors like to incorporate “Security Analysis” and “The Intelligent Investor” but I haven’t seen much consideration for “The Interpretation of Financial Statements”. Admittedly, some of the examples are probably out of date but I think that digging a little deeper in the financial statements may be a worthwhile endeavor. Coming up with true valuations maybe be right considering all of the liberty GAAP allows and what appears to be an SEC which doesn’t really seem to mind “creative accounting”.

4. Have the site’s authors read “It’s Earnings that Count” by Hewitt Heiserman, Jr.? It’s another book which looks closely at financial statements and valuations. Do the author’s feel that any of those concepts are worthwhile trying to incorporate?

As you can probably tell, I’m very skittish about investing and I often suffer from “Paralysis by Analysis”. I would love to find a somewhat simple, Graham based procedure which I felt comfortable “Pulling the trigger”.

Thanks for any comments,

david

Can you give some insight into your backtesting procedure?

I’m having a hard time finding some of the historical data you are referencing – can you list a source?

In terms of educational material, I can’t find any good books on the subject and the only websites I can find which really deal with this are actually software programs. Did you use one of these programs or can you recommend a good book on backtesting?

Thanks,

david

Simple and useful

Glad you like it!

Dear all,

I am wondering if there is not a mistake in the calculator’s formula:

var EPSmgA = (A*.33333)+(B*.26667)+(C*.2)+(D*.13333)+(E*.06667);

var EPSmgB = (E*.33333)+(F*.26667)+(G*.2)+(H*.13333)+(I*.06667);

It doesn’t take into consideration the 9th year (form J). Is that normal?

Thanks for the note. That year is actually not needed in the calculation of the value, and I’m not sure why it’s included in the form. I’ll try to take it out to make it.

The method says you shoudl take into account the past 10 years. So, the formula should be:

var EPSmgB = (F*.33333)+(G*.26667)+(H*.2)+(I*.13333)+(J*.06667);

and so taken into account the 9th year -> J.

The formula I developed for growth looks at the EPSmg for the current year and the EPSmg from 5 years ago, and uses the change in the two to estimate the growth going forward. The calculation for EPSmg from 5 years ago takes into account the EPS from 5 years ago, 6 years ago, 7 years ago, 8 years ago, and 9 years ago. It does not use the figure from 10 years ago, so the reference to needing the earnings from 10 years ago is mistaken. That I can easily change, but I’m having a little difficulty changing the coding of the form to remove the input box for that year of earnings.

Hello all,

After few research, I found a French company BIC (EPA:BB) which seams to be undervalued & Defensive.. Can anyone confirm my analysis?

I don’t currently evaluate foreign companies as they don’t translate very well into the spreadsheet I use, but I’m working on finding a solution.

I am still learning how to read your reports and understand your terms. I have a simple question that I did not see an answer for in your explanation of terms..Is the MG Value your forecasted price per share and what is the period that this guess good for?

Thank you

The MG Value is an estimate of the intrinsic value of the company. It can best be explained by imagining that you have an item that is worth $100. You know it is worth $100 because you have done research into the value. Consider $100 to be the intrinsic value. Meanwhile, every day someone comes to your door and offers to buy the item from you. Each day they offer a different price. Some days, the offer is to buy it from you for $75, other days they offer $125, and on a few occasions they offer $100. The offer they make in no way changes how much the item is worth, and sometimes it is downright astonishing that someone would make such an offer. That’s the same way the market works. You spend time researching a company and using its financial statements you determine its intrinsic value. That amount is not related to the stock price, but helps you decide whether the offer from the market is a good offer or not.

So if a company’s intrinsic value is $100, you know that if the market price is $50, it may be a good time to buy. Alternatively, if the market price is $200, it may be a good time to sell.

The problem is that the market is inherently unpredictable, so it is extremely difficult to predict when (or even if) the price will be close to the value.

So to answer the question, the MG Value is not a forecast in the traditional sense of the word, but one would expect that over a long time it is better to invest in companies trading for less than their value than to invest in companies trading for more than their value.

Why is your formula P/E * P/B < 50 where as Graham's original formula is P/E*P/B < 22? At first sight you seem to have been less cautious.

Great question. It is a little less conservative, and I must admit that it has been so long since I decided on this requirement, that I cannot remember exactly why I changed it. However, I think part of the reason is that the truly defensive investor of today invests solely in index funds, so an investor using Graham’s Defensive Investor requirements is a little bit less passive than Graham originally intended. As a result, the Defensive Investor of today may be willing to take on a slightly higher level of risk.

Although I understand you sold at 110% so that you had money to find more bargains, I don’t think Graham would advocate selling a good company bought at a good price even though it is currently overvalued? I really dig your website and only ask this question because I think you know better than I. Thanks

Fran – I’m not sure if I know any better than the next person, but my view is that you should not hold an investment if you would not buy it today. I always want my money to be invested in the greatest opportunity for profit. If a company is priced well above its intrinsic value, I wouldn’t buy it so why should I continue to hold it, especially if there are other opportunities that may bring a greater return on the investment?

Hi, I am new to this site so please forgive my ignorance. I am looking at your report on PEP. Admittedly, it is from 2009, but I am just trying to learn how you determine the EPSmg. Using your eps figures from the report & applying the sum of the years digits weighting, for 2004 your have an EPSmg of $1.95. My calculations are returning $2.15 as follows:

2000 – 1.42 x 1.0667=1.51

2001 – 1.33 x 1.133= 1.50

2002 – 1.67 x 1.2 = 2

2003 – 2.04 x 1.2667= 2.58

2004 – 2.41 x 1.333 = 3.20

I am assuming you add these figures and divide by 5 and that returns $2.15.

Could you please explain what I am doing wrong? Thank you.

Eleanor,

Thanks for the comment. Using the EPS figures you have listed, the calculation is as follows:

2004 – 2.41 x (5/15) = .7953

2003 – 2.04 x (4/15) = .544

2002 – 1.67 x (3/15) = .334

2001 – 1.33 x (2/15) = .1773

2000 – 1.42 x (1/15) = .0947

Then the sum of all of those is $1.95.

I hope that helps!

Can you explain when you decide to use the 2014 estimates in figuring the EPSmg and the value & when you do not use the estimates. I ask because today, 4/15/14, in your e-mail you highlighted HCP & COV. On COV you used the 2014 estimates, but on HCP you did not. Your value on HCP was $59.34 When I entered the 2014 estimate of $2.98 the value calculated as $81.51. Quite a difference!! Could you please explain the theory behind this. Thank you for this website. It is very informative!

Eleanor,

Great question! Here are my guidelines for using earnings (based on fiscal years):

If Q1 has been released, use the lowest available analyst estimate for the full year.

If Q2 has been released, use actuals for Q1 and Q2 and the lowest available analyst estimate for Q3 and Q4.

If Q3 has been released, use actuals for Q1, Q2, and Q3 and the lowest available analyst estimate for Q4.

If Q4 has been released, use actual figure for the full year.

In the case of COV, the 2014 Q1 earnings are available, so I used the lowest available analyst estimate for 2014, which is $3.95.

For HCP, the 2014 Q1 earnings are not yet available, so I did not use any estimate for 2014.

I hope that helps!