Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Kerim’s May 2014 Rankings
#1
So I’ve spent the past few weeks doing a pretty significant overhaul of my tracking and scoring spreadsheet, and thought some of you might be interested in the results. This shows the scoring for the 28 stocks that I follow most closely. I’ll be adding more soon, but these are the ones I care about most, and for which I have done a complete update in the past few weeks.

I also overhauled the scoring system, putting a little more weight on initial yield and adding a lot more granularity to some of the categories. This caused some of the scores to change quite a bit, but most of them stayed in the same ballpark.

The biggest surprise was with T, which has a much higher score than previously. I looked carefully at it, and the reason has much more to do with the fact that I had not updated T’s information in a while than it does with the revisions to my system. T’s earnings are in much better shape than a couple of years ago, and the payout ratio is back in “sane” territory.

Anyway, please take this with a huge grain of salt, though I will be happy if it generates lively discussion. I calculate many of these numbers in unconventional and arbitrary ways; the scoring is extremely arbitrary, reflecting my particular biases and whims; I make some completely subjective adjustments; and not all data is updated with the same frequency. Furthermore, this is only a stepping off point – this excludes huge swaths of relevant information, such as cash flow and balance sheet information.

Anyway, enjoy!

   
Reply
#2
OK. I'll ask the obvious first question:

How do you calculate your score?
=====

“While the dividend itself is merely a rearrangement of equity, over time it's more like owning an apple tree. The tree grows the apples back again and again and again, and the theoretical value of the tree doesn't change just because of when the apples are about to fall.” - earthtodan


Reply
#3
Nice!

Is the P/E TTM or forward?
Reply
#4
The scoring is a lot of complicated silliness -- I'll try to provide some insight when I have some time over the weekend. Gotta get the kid to bed now.

As for P/E, I use an average of current price to (1) 2013 earnings, (2) 2014 estimated earnings, and (3) ttm as reported by yahoo or morningstar.
Reply
#5
(05-02-2014, 08:29 PM)Kerim Wrote: As for P/E, I use an average of current price to (1) 2013 earnings, (2) 2014 estimated earnings, and (3) ttm as reported by yahoo or morningstar.

I can understand the average of ttm and fpe, but what's your thinking behind averaging in 2013 earnings? Seems to me it pulls things pretty far into the past.

Also curious to see how the gears work in the ranking machine.
Reply
#6
(05-02-2014, 11:36 PM)earthtodan Wrote: I can understand the average of ttm and fpe, but what's your thinking behind averaging in 2013 earnings? Seems to me it pulls things pretty far into the past.

Also curious to see how the gears work in the ranking machine.

Reasonable minds will differ about it, but I include 2013 earnings into the average because in my mind, it is the only truly "solid" number of the three. I'm not willing to spend the time to calculate a ttm earnings number myself, but I don't fully trust the ones reported by morningstar or yahoo either. I don't know if they are using diluted numbers, GAAP, non-GAAP, or what. Nailing down earnings is hard enough, but I am not comfortable relying on a reported number that I know little about. The forward projection of P/E based on expectations about 2014 earnings is valuable, but also based on guesswork. 2013 earnings, while arguably more and more stale over the course of 2014, at least are a solid foundation, and recent enough and relevant enough in my mind to make up a third of the average P/E number that I use.

I'll post a bit more in a few minutes about my "system" if I can.
Reply
#7
   

So here is a sample of the data I keep for each of the companies on my watch list -- MCD in this case.

The top half is pretty self-explanatory; I just compile earnings and dividend numbers by quarter. Numbers in green are already announced (or most likely given history); numbers in red are guesses.

In the bottom half, the single-cell boxes that have borders around them (there are 8 of them) I fill in manually. The rest of the numbers are calculated automatically from the other data.

In the big "scoring box" (cells J15 to M26), you can see how I assign points. There are 100 possible points. Dividend streak is worth 20 points; current yield is worth 20 points; payout ratio is worth 15 points; dividend growth and earnings growth are each worth 10 points. The business model category is subjective, but the basic idea is that KO has a way better / easier to understand / easier to execute / less overhead / less risks business model than, say, a tech firm or a financial firm. Basically, if I could imagine running the place myself, it gets a lot of points in this category; if I couldn't, it gets fewer.

Finally, there is an adjustment opportunity, that I use to tweak the outcome in certain situations. The best example I can think of is MO. The formula that assigns points for payout ratio is geared in such a way that points diminish pretty rapidly above a payout ratio of 60 percent. By 80 percent, the formula assigns almost no points. But MO specifically targets a payout ratio of 80 percent, so I don't feel that it is a fair assessment to give it no points in that category. So I use the adjustment box to make that "right."

I'll follow up with another post soon about the formulas for assigning points in each category, if I can get up the nerve. It is pretty arbitrary. But I like the idea of throwing it open to the group. With the group's feedback, maybe I can make it a whole lot better.
Reply
#8
Thanks for sharing this. I'm impressed with the amount of bookkeeping you're willing to take on. I'm also impressed that you're limiting yourself to 10 points for the subjective category.

How do you translate the score into a portfolio allocation?
Reply




Users browsing this thread: 2 Guest(s)