Investors are always looking for a system to achieve higher returns with less risk. How to find such a system? Why not start by looking at what one of the world’s greatests investors, Warren Buffett has been doing for many years? Buffett has said many times that he likes to buy companies with the following characteristics:

1. Simple businesses that he understands
2. that have predictable and proven earnings
3. with economic moat
4. which can be bought at reasonable prices.

The people at Gurufocus have tried to convert the last 3 characteristics in rules for investing, thus creating a system. They also conducted a back test study of parts of their system and also the system as a whole over the period between 1998 and 2008. In this post, I will give a summary of their publication. The original articles can be found here:

• Part I: Introduction of Predictability Rank
• Part II. Under-Valued Predictable Companies and Buffett-Munger Screener
• Intrinsic Value, Discounted Cash Flow and Margin of Safety
• What worked in the market from 1998-2008? Revisiting Business Predictability

Introduction

Buffett-Munger style investing means buying stocks of great companies at good prices. (Actually Buffett also invests in different ways, like buying companies outright, but this is not possible for the average investor. Also he invests in other securities like bonds, but this is not the focus of this post.) Good companies will continue to increase their earnings per share in the future and the stock value will eventually increase to reflect the growing earnings (in the long the stock market is a weighing machine). If the stock can be bought at good prices, you have now found a good investment. We are searching for a system to help us find Buffett-Munger type investments. The system should work in all markets to help us find investment which have a high potential return while simultaneously having a low risk for permanent loss of capital. That is, it should follow Buffett’s rule number 1: ‘Never lose money’. The 2008/2009 recession is a great period to test the system, since if the system contains fundamental flaws, these will probably be exposed in this crisis.

Predictability rank

When investing, you are buying a part of a company’s future, not it’s past. It’s very difficult to know the future, but would a consistent performance in the past be a good indicator? Gurufocus introduced the predictability rank: a ranking between 1 star (not predictable) and 5 stars (very predictable) based on how consistent a company’s revenue per share and EBITDA (Earning Before Interest, Tax, Depreciation, and Amortization) per share have been over the past 10 fiscal years. Any company that has ever had an operating loss, is considered unpredictable.

When you look for correlation between the predictability rank and investment returns between 1998 and 2008, you find that stocks of 5 star rated companies had an average annualized gain of 15.4%, while stocks of 1 star rated companies had an average annualized gain of just 6.7%. The average annualized gain of all stocks over the period is 8.4%. It is important to note that these results do not take into account valuation: stocks of more predictable did better during the period even if they were very expensive when they were bought.

It is also important that of the stocks of 1 star rated companies, 45% is still in a loss after 10 years (that is, the price per share at the end of the period is lower than at the start of the period). Thus, buy-and-hold does not work if you buy bad companies. Of the stocks of 5 star rated companies, only 3% is in a loss after 10 years.

The role of market valuation

So far we have seen that, regardless of valuation, more predictable companies have on average a higher return and a lower risk of loss over longer periods of time. What happens if we search for stocks of highly predictable companies which can be bought at good or great prices? This section will look at that question.

How do you know if the a stock is cheap, reasonably priced or expensive? The people at Gurufocus have chosen to use to price-earnings-to-growth ratio (PEPG-ratio) which they calculated by dividing the P/E-ratio by the average EBITDA growth rate of over the past 5 years. If you look at the investment returns of the 100 most predictable companies in relation to their valuation, you get the following data:

	Under-valued (PEPG < 1)	Fair-valued (1 < PEPG < 2)	Over-valued (PEPG > 2)
Total number of stocks	25	35	40
Annualized average gain	20.3 %	13.2 %	13.3 %

The S&P 500 gained just 2.7 % annually over the same period. Selecting undervalued predictable companies as opposed to the S&P500 would clearly give you a much higher return.

What about the probability of loss? For the highest 200 predictable companies which are also undervalued, the probability of loss if you hold the stock for 10 years is reduced to less than 4%. If you look at the average for all stocks, the probability of loss is 37% and if you just look at the unpredictable companies the probability of loss is 45%. A very large difference!

Conclusion: we already knew that stocks of highly predictable companies on average outperform stock of unpredictable companies. Now we know that when you buy undervalued stocks of highly predictable companies, the return is increased even further while simultaneously reducing the risk of loss.

Competitiveness of business

How do you know if a company has an economic “moat”? The company’s gross profit would seem to be a reasonably proxy. If a company can maintain or even increase it’s profit margin while growing it’s business, it is said to have a moat.

The Buffett-Munger screener

The people at Gurufocus have developed from their research, a stock screener which they call the ‘Buffett-Munger screener’. The stocks that can make it into this screener are:

1. Companies that have high predictability rank: they can consistently grow their revenue and earnings.
2. Companies that have competitive advantages (an economic “moat”): they can maintain or even expand their profit margin while growing their business.
3. Companies that incur little debt while growing their business: safe companies (as defined by their debt-to-equity ratio?).
4. Companies that are fair-valued or under-valued (can be bought at fair prices): the price-earning-to-growth ratio is smaller than 2.

Taking into account the 2008/2009 crisis

All above data mentioned above was for the period between January 2nd of 1998 and Augustus 31st of 2008. If you extend the period to March 25th of 2009, thus including part of the recession of 2008-2009, it can be seen that the top 100 most predictable companies had an average annualized gain of 12.2% while the S&P 500 had an average annualized gain of -1.1%. Thus, investing in predictable companies is safer (you still made money during the period) and more recession proof than investing in unpredictable companies.

Personal notes

Is the Buffett-Munger screener a perfect copy of what Warren Buffett would do? No, since the system will miss securities which would be clear buys for Buffett, and it might also let pass through securities which Buffett would never buy. For instance, Coca Cola, a company of which Buffett has said many times it is great, only has a predictability rating of 2.5 stars and thus would not pass through the Buffett-Munger screener. Wells Fargo and Company, another large investment of Buffett, only has a 1 star rating.

That said, I think the system is a very good starting point, and if you follow it, you might expect to do a lot better than the S&P. If you can enhance the strategy by using the system as a starting point and then adding to it by using your own knowledge, experience, skills, intelligence, etc. to find the winners and avoid the losers, it would be even better.