Trading with Benjamin Graham's formula

Back to Community

Trading with Benjamin Graham's formula - Fundamentals and bonds

edited Jun 12, 2020

The Benjamin Graham formula is a formula proposed by investor and professor of Columbia University, Benjamin Graham, often referred to as the "father of value investing"[1]. Published in his book, The Intelligent Investor, Graham devised the formula for lay investors to help them model growth formulas in vogue at the time of the formula's publication. - Wiki

Also I have combined with the bonds trading strategy from Chris Cain's recent post 'New Strategy - Presenting the “Quality Companies in an Uptrend” Model'

Formula : Graham = EPS * (8.5+2g)
Buy: Graham - Market Cap > 0
which is when the intrinsic value is great than current value

Flaws:
1. The period that I chose performed best, before or after doesn't look good.
2. As in the original book, g should be the 'reasonably expected 7 to 10 year growth rate', where I would use two years mean of revenue growth rate. Here, I do not know how to use past Fundamental data so I simply picked 'revenue_growth.last'

Possible improvements:
Use Revised formula: EPS * (8.5+2g) *4.4 / Y

Question: How to use Historical Fundamental Data?
As I searched in quantopian forum, the answer i found was its not available. Well, please, please let me know if you know how. Thanks!

5 responses

Tentor Testivis

Jun 12, 2020

Simply leave out .latest in the factor inputs, choose a window of two years and compute the mean of revenue_growth. I assume the other values should be the most recent?
Here a backtest, the performance looks worse though...

ZT YE

Jun 12, 2020

@Tentor Testivis, Thanks a lot! Its interesting that it would perform worse, good learning experience though.

Vladimir

Jun 12, 2020

Performance metrics will be much worse if you include both market crashes in backtest.

ZT YE

Jun 13, 2020

@Vladimir, it doesn't perform that well in bad times. I might try adding a trending following strategy see if it will help.

Patrick Marum

Sep 21, 2020

I replaced your formula with the following code:
The idea I isolated the price from the basic equity value formula divided by the share.
equity value = enterprise value - debt
intrasic value = equity value / shares

class Intrinsic(CustomFactor):  
    inputs = [Fundamentals.enterprise_value,  
              Fundamentals.total_debt,  
              Fundamentals.shares_outstanding]  
    window_length = 260 # 252 trading days per year  
    def compute(self, today, assets, out, enterprise_value, total_debt, shares_outstanding):

        out[:] = ((enterprise_value[-1] - total_debt[-1] ) / shares_outstanding[-1] )

However, the results are better but it is not glorious.

You've successfully submitted a support ticket.

Our support team will be in touch soon.