By Robert Barker Can two guys -- one on each coast -- plus a computer model run a successful tech-stock fund? That's what John Gipson and Mike Cohen are trying to prove with their Alpha Analytics Digital Future Fund. It got off to a great start last year before losing 37% in the fourth quarter, enough to put it 13.9% under water for the year. The horror of it is, in a year when the average tech fund lost 33% (according to Morningstar), Alpha still did better than 92% of its rivals.
And it's bouncing back, up 19.6% through Jan. 18, compared to 9.5% for the average tech fund. Unlike many funds, which use computers as a first step to sift through a bunch of potential stock picks, Gipson and Cohen save the computer analysis for last. Lately, the model has been pointing to Intel (INTC), among other large-cap names. (With just $3 million in assets, the no-load fund is too small yet to have a ticker symbol. Its CUSIP is 02071R308.)
To learn more, I spoke the other day via conference call with Gipson in his Silver Spring (Md.) office and with Cohen in Pleasanton, Calif. Here are edited excerpts of our conversation:
Q: How come there are two of you to run a $3 million fund?
Cohen: I'm the long-term strategic manager, and John is the short-term, or tactical, manager. I focus on the long-term fundamentals of companies. I'm out here based near Silicon Valley, [so] I go to a lot of trade shows, [and] I talk to a lot of engineers. I try to identify themes in technology that I think are going to be significant. And then I try to find the best companies that are going to capture those themes.
Q: What does John do?
Cohen: John takes a proprietary quantitative model, which tries to predict performance one month going forward, and then he determines the weightings of the fund -- where we're going to put new money.
Q: Give me an example of a stock that did well for you last year. How did you identify it, and how did you trade it?
Cohen: Newport (NEWP) is an excellent example. Newport is one we bought back in January. At that time, all fiber-optics prices were sky-high, and I was looking for a way of expanding into a [fiber-optic stock] that really wasn't discovered yet.
Q: Go on.
Cohen: JDS-Uniphase (JDSU) said it was going to ramp production as hard and fast as it could, promising a doubling or tripling of manufacturing capacity over the following year. So I tried to identify companies that I though might benefit from helping to automate the manufacturing of optics. I looked at companies like Veeco (VECO) and Newport, and after looking at the companies, Newport seemed the clear pick.
Q: What next?
Cohen: Then, I went to a trade show called Photonics West 2000 out in San Jose. I talked with [Newport] at the booth, and by talking with them and by listening to conference calls, I secured that JDS-Uniphase, SDL (SDLI), and E-TEK E-TEK
were customers of theirs. And these are all companies that are constrained by manufacturing capacity, not demand.
Q: So once you bought Newport, did you hold steadily or did you trade it?
Cohen: We pretty much held steadily, and then when it was about $150, we thinned some of our winnings, but we didn't trade it in and out. And it has been very volatile. Most of our companies have been very volatile.
Q: How many issues do you typically hold?
Cohen: We set a target of 25 to 35 companies. I believe we have 31 or 32 companies [right now].... Our biggest position is Newport.
Q: And how much of the portfolio is in Newport?
Q: That's a lot. What's a stock that you recently added to the portfolio?
Cohen: Network Appliance (NTAP).
Q: Refresh my memory: What does the company do?
Cohen: Network-attached storage. We were heavily in the [storage-area network] companies as a sector theme, picks like Emulex (ELMX), Q Logic (QLGC), EMC (EMC).... But the real leader in [network-attached storage] is Network Appliance.
Q: What's your average cost in Network Appliance?
Gipson: Our average cost was about $63 a share.
Q: What else do you own?
Cohen: Our top five, in order, [are] Newport, Emulex, Q Logic, Applied Micro Circuits (AMCC), and SDLI
Q: What do AMCC and SDLI do?
Cohen: SDL is a laser company. They're being purchased by JDS-Uniphase, which is a company we also own. Both of those have been holdings since the inception of the portfolio.
Q: So will you get JDS-Uniphase shares when that deal closes?
Cohen: We'll have shares in the combined company.
Q: And you're planning on holding on to it?
Cohen: Most likely, yes. The deal is supposed to get shareholder approval Jan. 26.
Q: And what about Applied Micro Circuits (AMCC)?
Cohen: AMCC is a semiconductor company. Whereas lasers initiate the beam in optics, on the other side you actually have to have something that detects the light and converts it back into electricity. These are receivers or transceivers, and AMCC is a major player in that market.
Q: John, what is it you do once Mike comes up with names of stocks to buy?
Gipson: I have a proprietary model that has 50 parameters in it. This model looks at what the stock market has valued over the last nine months, roughly speaking. The rationale behind it is that at different times, different characteristics come in and out of favor.
For example, sometimes large-cap stocks are in favor. Sometimes small-caps, sometimes stocks that have growth characteristics. Sometimes value. And we don't try to predict what's going to happen so much as to follow what has happened.
Q: So underlying that is the assumption that past patterns will repeat?
Gipson: Past patterns will repeat, and the other crucial factor is that the market is changing, so you can't do something that is static.
Q: What led you to this?
Gipson: I have a PhD in physics from Yale University, and I worked at NASA for many years. And my brother, Reg Gipson, who is the president of the company, wanted me to do some investigation into ways of trading, to see if there were ways that you could have an edge.
Q: Does it work?
Gipson: What we've found is that on the sell side it has been right about 70% of the time and wrong about 15% of the time. On the buy side, it has been right about 60% of the time and wrong about 30% of the time. It's not always successful, but you don't have to always be successful.
Q: What's a success story?
Gipson: A year ago, the model was recommending against large-cap stocks, like Microsoft (MSFT) and Texas Instruments (TXN) and Lucent (LU), so we kept out of those names. Nowadays, the model is starting to like some of those names again, so we've gotten back into Microsoft and put money into Texas Instruments.
Q: Mike, what areas of technology are you staying away from?
Cohen: We've stayed away from the fixed wireless. However, we do love third-generation [3G] wireless.
Q: How are you playing 3G?
Cohen: Our favorite pick there is Qualcomm (QCOM), largely because we believe [cell-phone protocol] CDMA is the dominant technology, in both of its flavors, CDMA 2000 and wide-band CDMA. Qualcomm will get royalties either way. However if it moves to CDMA 2000, they're definitely better off.
Q: What's wrong with fixed wireless?
Cohen: We see fixed wireless as not having a good reward-to-risk ratio, largely because of another potentially disruptive technology out there. There are companies, TeraBeam Networks and AirFibre, that would be potentially using lasers.
Q: Those are private companies, right?
Cohen: Those are both private, [but] AirFibre is partnered with Nortel (NT) and TeraBeam is partnered with Lucent. [These] major partners could bring them to market quickly if the technology is ready. So we see that as potentially disruptive to the entire industry. Also, we shoot for something that's proprietary. Fixed wireless is essentially microwave technology that has been around since the end of World War II and most of the patents there have expired. If there were large profits to be made, startups could enter the space easily, eroding what other companies have captured.
Q: How often do you guys talk?
Cohen: Every day, and usually several times a day, either by e-mail or phone. Barker covers personal finance in his Barker Portfolio column for Business Week. His barker.online column appears every Friday, only on BW Online.
Questions? Comments? Let us know at our barker.online interactive