Sep 27, 1999 at 12:00AM
I proposed a scoring system and decided to back-test it this week. Without going into the nastiness of the math, I checked for statistical correlation for my scoring system and got nothing. I tried desperately to come up with some sort of a scoring system that would give results, but nothing came up as statistically valid. Looking at 1994 financials, I tried to correlate them with future returns and found nothing that gave meaningful results. I did have more success if I dropped out a few "troublesome" companies -- but is that a proper way to come up with a system?
Mark Twain (one of Fremont's contemporaries) put it quite well: "There are three types of lies -- lies, damn lies, and statistics." I suspect that if it really were possible to screen a company on the basis of blind data analysis, the mutual funds would have picked it up long ago. They could select stocks by computer, fire the high-priced help, and beat the market year after year. Some funds may select stocks by computer, but they haven't done the other two successfully.
So, where do we go from here? I sorted the companies by the returns they provided in the past five years, and the winners did have one thing in common. They were not phone companies, they provided equipment. (See the results on my website, http://www.mindspring.net/~grunkle/returns.htm.) Motorola (NYSE: MOT) was an exception to this. Worse yet, in 1994, all of its numbers looked very good. In five years, it returned a total of 75%, or 12% annualized. This was respectable, but well below the market during that period. However, there were some problems Motorola had with Iridium that could not have been foreseen. Also, they stuck with analog cellphones when the world wanted digital. Few things can be predicted, which is another reason for diversification.
After eliminating the telephone companies, we have eight left: Intel (NYSE: INTC), Compaq (NYSE: CPQ), Hewlett-Packard (NYSE: HWP), Nortel (NYSE: NT), Motorola, Lucent (NYSE: LU), Scientific Atlanta (NYSE: SFA), and IBM (NYSE: IBM). How do we narrow down the list from here? I'm going to chuck all my numerical analyses and use the way I've successfully chosen Drips in the past. Here goes:
Intel -- I loved Andy Grove's book Only the Paranoid Survive, and I won't buy a computer with any other processor. I'm suspicious of what other corners a computer maker may have cut if a different processor is used. So, Intel stays.
Compaq -- I've had three Compaq computers, two of them were only reliable in crashing frequently. Customer service was impossible to contact. It is eliminated.
Hewlett-Packard -- I have a Hewlett-Packard printer and I love it. My son has a Hewlett-Packard computer and loves it. I love their calculators, their faxes, and everything they make. It stays.
Nortel -- Don't know anything about them, they go.
Lucent -- I live up I-85 from Lucent. I see their employees walking around neighborhoods and shopping malls wearing shirts with the company logo. That's a good sign. I never did that when I worked for a big company. They stay. Oh, they also have Bell Labs, a major source of innovation in telecommunications.
Motorola -- They really blew it with Iridium, and their insistence on staying with analog instead of digital cellphones cost them market share. It's too much to come back from, so I will pass on them.
Scientific Atlanta -- My wife and I are arguing about this right now. I refuse to include them as a possible investment candidate because I know little about the company. She's (at this moment) getting the business section of the paper to show me an article about them. She says they hit a new high, which is all you need to know. I don't agree. Whom do you side with?
IBM -- IBM has been around a long time, and their stock price has done well recently. However, right or wrong, I don't feel comfortable investing in them. They have never impressed me as an innovative company. Can one of you out there persuade me to feel differently?
This leaves Intel, Lucent, and Hewlett Packard. As an engineer, I'd feel better if there were a surefire way to come up with companies to invest in by computer screening. However, I think Peter Lynch is right. It is best to invest in what you know. I'd like to wrap this all up next week, but please let me know what you think by e-mail or on the Drip Companies board. If you are interested in seeing where I got my data for 1994, check out Value Line, which you can get it at your local library.
- Sep 27, 1999 at 12:00AM