I have certain thoughts on a couple of Graham formulas that are widely used by value investing folks, especially for the Indian markets. Please feel free to post your comments to correct my understanding.

a) __Graham’s Formula__

I have seen multiple investors use this formula quite blindly and call a stock undervalued/overvalued depending on whether the Graham price derived from this formula is greater than or lesser than the current market price. Graham published this formula in 1974 and took only US data. The formula per se states

Price = (EPS* (8.5+2*g)*4.4)/AAA rate

where,

EPS = TTM EPS for the past 12 months

8.5 = P/E multiple for a no-growth company

g = the company’s long term earnings growth estimate (5 years)

AAA rate = current prevailing AAA bond rate

4.4 = The average yield of high-grade corporate bonds in 1962.

For Indian markets, I guess we can use the formula as-is, but for the number 4.4. Why do you want to use 4.4 when the average yield of high grade corporate bonds in India is around 10-14% currently? I have no idea why people use 4.4 here. If you do, please let me know.

I will tell you my approach here when I use Graham’s number.

EPS = Average (median) EPS over the past 5 years (10 years, if data is available)

7 = P/E multiple for a no-growth company

g = 25% of the past 5 year CAGR (I am pessimistic that way, and I can’t predict future growth)

AAA rate = currently use 10%

12.5 = Avg. yield of high-grade corporate bonds currently

So, the modified formula would work out to

Price = (Avg. EPS * (7+1.5*(25% of 5 yr CAGR)) * 12.5)/10

What do you think of this approach? I have used the logic of Graham, but not the 4.4 multiple that Graham used. Does it lead to over-stating the Graham price?

b) __Graham Number__

This formula is also widely used, although in my experience in Indian markets, the resultant values are a bit on the higher side.

Price = Sqrt (22.5 * Book value * EPS)

Let’s not use this formula blindly. Let’s think about – How did this formula come about?, How did he arrive at the number of 22.5? etc.

Graham thought that nobody should be paying more than P/B = 1.5 and P/E = 15 for a stock. How did he arrive at the number ‘15’ for P/E? Well, he thought that nobody should be willing to pay more than the AAA bond yield at that time. AAA bond yield at that time was 7.5%. Therefore, AAA P/E will be 1/7.5 ~ 13.3, rounded up to 15.

So, Price/Book * Price/Earnings = 1.5*15

Price (squared) = 22.5 * Book value * Earnings

Price = Sqrt (22.5 * Book value * Earnings)

Now that we know how the formula was derived, let’s modify it to Indian standards.

Let’s not have P/B more than 1.25. Current AAA bond yield is 10%. Therefore, AAA P/E will be 1/10% = 10. That is we should not be willing to pay more than P/E = 10 for any stock. That modifies our formula to,

Price (squared) = 1.25*10*Book value * Earnings

Price = Sqrt (12.5* Book value * Earnings)

Again, for conservative calculations, I use average of the past 5 years (or 10 years) of data for Book value and Earnings.

The one grouse that I do have with this formula is that for commodity stocks, the Indian stock markets rarely value companies at more than 8-10 P/E. So, using this formula indiscriminately will lead to poor investments. Do look into company details, capital structure etc. before investing.

c) __Simpler Graham__

Some other investors screen their companies based on an interview by Graham in 1976 titled “The Simplest Way to Select Bargain Stocks” where he stated that

i) P/E should be not be more than 7x-10x

ii) Equity/Assets ratio > 0.5 (in other words, D/E < 1)

Again, P/E screener should be changed for the Indian scenario. Graham stated that the yield on stocks should be atleast 2 times the AAA bond yield (to compensate for the risk). Back then, the rates were 7%. So, the yield on stocks need to be 14%, which in turn means P/E not more than 7. If rates dip to 5%, then P/E not more than 10.

In the Indian scenario, AAA bond yield is 10%. Stock yield should be 2*10% = 20%; and therefore stock P/E should not be more than 1/20% = 5. So, for the Indian scenario, the rules change to

i) P/E should not be more than 5x-7x

ii) Equity/Assets ratio > 0.5

**P.S: **Of course, one of the biggest rule of Graham was diversification. He used these quantitative formulas on a portfolio of 30 stocks and invested in them. Some turned out to be multi-baggers and some duds, but as a portfolio beat the market. So, if you are trying to use these formulas for investing in one or two stocks, kindly do due diligence and use these formulas with discretion.

#1 by Suhail on November 15, 2011 - 3:48 AM

Couple of elementary questions:

1) Where do I get a quote for typical ‘AAA’ bond and if possible historical values? I mean I understand it’s the bond rate of individual companies but is there a place where I can pull up their quotes? Especially inside my online trading acct. Or is there a shorthand thumb-rule which defines a typical AAA bond rate as function of some RBI rate?

2) For EPS – you mean ‘median’ or ‘mean’? Arithmetic avg is the mean (sum of values/total), whereas median is the point at which exactly half the values are above and half below. Or am I getting it wrong?

3) With yr revised calculation the PE < 5 criteria sounds astonishing. Would love to know if you've discovered any good quality companies (satisfying other general criterias from yr screening list) with PE 8-9. And if it’s a relatively talked-about stock (say nesco) it’s even higher. I know we don’t have to use PE exclusively, and with all usual disclaimers attached, but still wondering if you’ve found anything worthy in yr portfolio where PE~5?

Thanks.

#2 by Suhail on November 15, 2011 - 4:02 AM

whatay #3, correction: It should read:

“Most value stock ideas I read about are in PE 8-9 range. And if it’s a…..”

#3 by Kiran on November 15, 2011 - 11:46 AM

1. SBI had issued AAA bonds at 9.5% last year. Now, with increase in interest rates and all that jazz, I take it as 10%. If that’s a confusion, you can use 10 year Govt. bond yield (currently 9%). You can find history of these yields in any finance site.

2. You are absolutely right. I usually use the median because its a better bet. However, for a set of values which don’t range very widely, mean is approx. the same as median. Only when you have extreme values within a set, mean diverges widely from the median.

3. It helps to recollect that Graham was more of a ‘quantitative bargain’ guy along with ‘diversification’ guy with a time frame of 3-5 years. As I stated in my P.S, a bunch of stocks with PE < 5 would work out better than any single stock with PE<5 using Graham's criteria. PE < 5 stocks are usually beaten down severely by the market due to some issue or are commodity companies where cycle turns once every 3-5 years. So, Graham was basically better on 'reversion to the mean' logic for the former, and commodity cycle for the latter to gain benefits and beat the market ultimately over a 3-5 year period. Selectively choosing Graham rules will lead to poor investments.

#4 by neeraj on November 16, 2011 - 2:47 AM

Really nice post Sirji..not accepting the ‘accepted’ formulas and going into their logic is a great thought process to have.

Cheers!

Neeraj

#5 by Kiran on November 16, 2011 - 8:07 PM

@Neeraj – Thanks Sir!

#6 by

Manishon November 17, 2011 - 2:13 PMReally a thought provoking post!

#7 by

Vijayon November 17, 2011 - 9:32 PMImpressive

#8 by

Ashish Patelon November 20, 2011 - 1:25 AMNice learnt some good things from your post 🙂 I only knew one formula till now 😀

#9 by

Sujal Shethon November 21, 2011 - 2:08 AMHi, Your article is really informative and innovative. I’m also currently reading graham’s formula book and wondering that whether formula still applies to current time and also to indian market. Your article boost my confidence. But now with your formula, if i investigate the Tata steel, Current price is 400Rs. So if we take only 5% growth rate and EPS (avg. of last 5 years) 66 then it value (according to formula Price = (Avg. EPS * (7+1.5*(25% of 5 yr CAGR)) * 12.5)/10) comes to 1196 but as mentioned earlier, it’s current price is 400Rs. Why it is too below ? As graham also told that price shouldn’t be more than 20 times in last year earning or 25 times in last 10 years earning. Current earning is 70 so with that even price shouldn’t be more than 1400. Even we take conservative approach (15 times) then it should be 1050Rs. I don’t understand why there is too much difference in market price and graham formula price ? Am i missing something ? Please help me to sort out. If i’m right then is it good time to invest in Tata Steel ? 😉 🙂 Please reply me.

#10 by Kiran on November 21, 2011 - 4:55 PM

@Manish @Vijay @Ashish – Thanks!

@Sujal – Never investigated Tata Steel Sujal, so don’t have a view. However, as I reiterated in my post, do not use Graham’s formula in isolation.

You can read up Rohit Chauhan’s post on Tata Steel…he had some concerns on contingent liabilities…

#11 by Dhwanil Desai on February 29, 2012 - 10:14 PM

Hi Kiran,

A great post indeed. You have deciphered the logic behind the formula in most rational fashion. However,you have rightly mentioned that Mr.Graham suggested sufficient diversification to protect portfolio against downside from non-performers or money losers. I admire Mr. Graham’s framework very much, but in my humble opinion, many of his disciples have done a great job by factoring in qualitative aspects such as business economics, moat and management quality in addition to quantitative framework provided by Mr. Graham. Anyway, I think yours posts here are of superior quality.

#12 by kdaaku on March 1, 2012 - 2:05 AM

@Dhwanil – Hey, thanks for dropping by and your comment.

You have got a great blog too. Added to my blogroll. Let’s share ideas some time.

#13 by Dhwanil Desai on March 1, 2012 - 1:01 PM

@kiran

sure. we will keep in touch and share ideas/concepts.

#14 by Deepak Kumar on November 3, 2013 - 12:52 PM

1. Use Bond interest rate of 9.5

2. India’s 10 year current yield of aaa corp bond is at 8.56

3. don’t use 25% of cagr it will create problem use 2G instead, (2 X avg. growth rate of company).

how to get avg. eps, for this i take last 5 year data and calculate each year’s growth over the last one, this will give me 4 different growth number and I use average number of that as a company’s actual growth rate.

=(New Vlaue-Old Value)/Old Value*100 this will give you percent change in years.

=(value 1st year+value 2nd year+N+N+N+)/5 this will give you average growth rate of that company.

add: use cagr of the company to verify your growth rate , if the difference if higher you might need to adjust your inputs.

==100*((Last Value/Old Value)^(1/Noofyears)-1)

#15 by

mayankon March 1, 2015 - 11:50 AMsir u took [g= 25% of 5 yr. cagr] but as per my calculation we should take 5 or 10 yr. eps growth.

#16 by Kerri on May 17, 2016 - 2:56 AM

Got this website reference from Maru Gujarat Official Site. I want some useful syllabus for the preparation of airforce bharti. Also I want latest gujarat rozgaar samachar please give link

#17 by unnatib1981 on June 14, 2016 - 3:12 PM

What if the P/E of a stock (latest price / latest annual EPS) is over 50x and latest price to total assets is at 0.29x. Then is it still a good BUY??