Today I want to share with you a model portfolio that reinforces the concept of constant learning.
I spent a lot of time in 2015 talking with Tobias Carlisle about valuation. He has pretty much convinced me, with the weight of evidence, that enterprise value-to-earnings before interest and taxes is as good -- and sometimes better -- than price-to-book value as a measure of corporate value.
If you have not read Carlisle's books Deep Value and Quantitative Value (with Wesley Gray as co-author) I suggest you do so before the year gets much older. I also spoke with Gray on the subject late last year and he does a good job of reinforcing the point.
I have started using the EV/EBIT measure in many of my tests and searches and found it to be a valuable tool. After several discussions with Carlisle and Gray, they are adamant that it works as well with financial names as it does with other stocks. Personally, I think that P/BV remains the best way to evaluate financials and REITs, but as both Carlisle and Gray are quite a bit smarter than me, I included financials in my research and testing.
One screen in particular has outstanding results when using the EV/EBIT ratio.
I was trying to find a suitable approach to growth and income investing that could be used by more conservative investors who still wanted decent returns. I simply looked for names with EV/EBIT ratios of less than 6 and that paid a dividend. Then, I eliminated those with market caps below $100 million and further limited the list to profitable companies.
To avoid yield traps, I set an upside dividend yield of 7%, excluding any stocks yielding more than that. Finally, as a safety check, I limited the portfolio to only companies with F-scores over 5. I then bought the 50 highest-yielding stocks and rebalanced the portfolio every six months.
The results were so consistent that I went back and spent hours swilling coffee and checking the results by hand.
This approach beat the market over all the timeframes I tested (1, 3, 5, 10 and 15 years). Only those portfolios formed in late 2007 though late 2008 in the rolling back test had any serious declines.
Some may accuse me of data mining when running these tests, but consider the results here. These are profitable, dividend-paying companies with some margin of safety on the balance sheet. They are undervalued using the favorite metric of private equity and leveraged buyout shops. If finding a portfolio of stocks with these characteristics is data mining or curve fitting then pass me a shovel and a protractor please.
The model stays fully invested most of the time. Only in 2001-2003 and the period between 2005 and 2006 were we unable to find enough stocks to become fully invested. Even then, the model rarely goes above 30% cash. This approach did a little better than the market in the rough patch of 2008, but it did not escape the carnage entirely so it is best used by long-term investors willing to ride out the occasional bad spot.
The portfolio recovers much quicker than the market, and in similar fashion to other value models outperforms by a significant amount in the months following market bottoms.
As you can see in the spreadsheet, above, the portfolio is fully invested as we start 2016. It has a larger cap flavor than my usual output. The average market cap of $8.6 billion is somewhat skewed by super-caps such as JPMorgan Chase (JPM), Prudential (PRU), Valero Energy (VLO), Marathon Petroleum (MPC) and HP Inc. (HPQ), but the median market cap is still a healthy $855 million.
The average enterprise multiple is 4.2. The average P/E ratio of 14.7 may look high at first glance but when you adjust for the large goodwill adjustment at Stewart Information Services (STC) and substitute operating earnings, the average multiple drops to 14.
The average F-score is 6.5 and the average yield is 3.4% right now. The portfolio is pretty widely diversified across industry groups, with 18 different sectors represented in the mix.
Even if you are not a fan of quantitative approaches or think that back-testing is a bunch of hooey you would be mistaken not to use quant-derived lists for ideas and shopping list.
My list using enterprise value growth and income factors has some very interesting stocks with the potential for high returns in 2016.