Ever since I read Joel Greenblatt's The Little Book That Beats the Market, I've been thinking about ways to integrate his system into my own. The problem was, that Greenblatt's so-called "Magic Formula" does not take into account dividends at all, while mine certainly does.
For those of you who are not familiar with the magic formula, here's a quick summary: Greenblatt wants to buy excellent companies (i.e. ones that have high pre-tax return on capital) at a cheap price, which he measures by pre-tax earnings-yield - the higher the better.
I like this value-oriented approach a lot. But I can't use it as it is. Many of the companies that have a good aggregate score in these two criteria don't pay a dividend at all. Others pay very little.
One thing I could do, is change the magic formula, replacing earnings yield with dividend yield. But Greenblatt himself warns against tinkering with the formula in this fashion. The stocks I end up with may have exceptionally high dividend yield but no more than average earnings yield.
Instead I decided to look at a subset of the stocks recommended by the magic formula, that also satisfy the additional criteria of dividend yield. To achieve this I started out with the top 50 magic formula stocks over $2 billion in market cap (available on the Magic Formula Investing site). I sorted the list by dividend yield, in descending order, and removed the companies that pay a smaller dividend yield than the S&P 500 average of 1.8%. The results of this screen are presented in the table below.
If we exclude the top two entries, which are actually investment trusts, we get a list of ideas for stocks that fit the magic formula criteria, and also offer an above average dividend yield. Incidentally, the screen turned up two stocks I already own.
The stock picks generated by the screen should be considered on an individual basis. One advantage they certainly have is that they are not financials (the magic formula doesn't do financials). This means throwing them in your dividend portfolio will probably improve diversification, as I've noted in a previous post.
Full Disclosure: at the time of writing, the author was long BVF, UST, and XLE, which is in large part XOM