Institutional investors have powerful tools at their disposal to screen through reams of data. Part of the institutional investment process entails screening through thousands of securities looking for a needle in a haystack — stocks that fit certain investment criteria. From thousands of stocks, analysts can filter through a couple of hundred that fit these so called screens. With a couple of hundred stocks in hand, analysts set out to do the hard work analyzing these companies, comparing them to one another, speaking to management and whatever else hedge fund and mutual fund 
analysts do when looking at prospective investments.
If I’m a value investor, I’m probably going to use some metrics that focus on Return on Capital (RoC) or Return on Equity (RoE) and Earnings Yield (E/P). Growth investors might like to compare the price/earnings ratio (P/E) to the actual growth prospects of the stock in question. There are literally thousands of things to look at. So, where to start?
As we discussed in the “Piggybacking the Pros“, the Internet is a wonderful place to get information on stocks. From monitoring actual moves that uberinvestors make to seeing what your peers think of certain firms, the information is out there. You just need to be able to find it. Investment Screening 2.0 is all about using screening criteria from super investors like Warren Buffett to find the next big stock. Think of it as tapping the world’s greatest minds to see what they think of a given investment.
John Reese, a well-regarded entrepreneur with degrees from the two best schools in Cambridge, MA, has created a premium service that is the product of many years of research into the strategies of super successful investment managers. Having devoured numerous investment books and interviews with the world’s best investors, Reese embarked on an ambitious project to computerize this information. His firm, Validea, has created an algorithm capable of analyzing thousands of stocks according to differing investment strategies ranging from Fidelity’s Peter Lynch to Berkshire Hathaway’s Warren Buffett and applies screens to the overall market to find stocks that would fit these uberinvestors’ criteria.
John Reese joins New Rules of Investing for an exclusive interview.
John, can you tell us a bit about yourself?
John Reese, Founder, Validea: I have a degree in electronics and computers from MIT and a Harvard MBA. Professionally, I worked for TI (Texas Instruments) in software and in strategic planning for GTE and Verizon. I then founded my own company in computer networking and sold the company to GE Capital. Post sale, I was looking to invest my own funds and wanted to do it in a very methodical way. I began reading probably 50 different periodicals including investment magazines, news, and investment newsletters. I was faced with numerous opinions. Who should I listen to? Who has the best track record?
In 1996, I began tracking recommendations from all the media sources I was monitoring (TV, magazines), much like Mark Hulbert does to find leaders in the investment newsletter space. I found that I was still not getting an answer as to who had the best track record or strategy.
Things became clearer when I read One Up on Wall Street, by Fidelity’s Peter Lynch, an investor with a proven track record, who shared in a step-by-step manner his strategy of stock picking. I realized that this could be computerized and I actually programmed the Lynch guru, an AI model that could receive the data for a particular stock and analyze it according to pre-programmed Lynch criteria to see if it would past muster.
I then tie the guru models into a Reuters database (using more than 300 different models, per stock per day), only looking at US stocks and ADRs.
How much of John Reese is in the formula?
JR: I try extremely hard to mimic the guru’s strategy as closely as I can through real-life examples he’s written or spoken about. Sometimes, it requires interpretation. For example, in particular cases where the professional investor is not clear enough with his methodology. The Guru has to have clearly described his investment process with quantitative aspects to it. Peter Lynch, Benjamin Graham from Intelligent Investor, Dreman, Zweig, Motley Fool for Small Caps (written in book), O’Shaughnessy (45 years of what worked on Wall Street), John Neff (Windsor), and more recently, Joel Greenblatt, have all done this.
What does the Validea output look like?
JR: At a high level, Validea either summarizes a Guru score for a particular stock or we can drill down into an individual strategy, which criteria are employed by the particular guru and how certain investments meet or don’t meet this criteria.
In 2003, we came up with methodology that displayed a top 10 stocks/top 20 stocks portfolio generated by each guru out of a database of 6000 stocks. The system makes a recommendation every 4 weeks and the results are available every day.
We distribute this information via a subscription newsletter called, Hot List, which uses a consensus strategy of gurus, which basically lets each guru strategy vote on each stock and weights each vote according to the risk-adjusted rating on each stock, our website, various reselling packages we have with various brokerage firms, and packaged onto Yahoo Finance.
How do investors encounter your research? Where do they find you?
JR: It happens in three ways. People frequently find out about Validea because the media likes to write about us given our unique approach. We’ve run a contest via Microsoft’s Strategy Lab. Lastly, people get to us from a wide variety of sources, including a lot of Google searches for gurus. For example, investors read about Peter Lynch and all of a sudden, they now have a step-by-step way to follow his strategy or others like his. There are a lot of people who have read these classic investment books.
How are you seeing investors interface with your products?
JR: It used to be that prior to 2003, the detailed guru pages (ie., the Ken Fisher model or the Ben Graham model) were regularly the most popular pages of the site. Once we began offering the models, those became the most popular. We began to hear from readers — some were tracking us for a long period of time, others were running their own portfolios, and some percentage of these actually wanted us to take over management of their portfolios. We became an RIA (Registered Investment Advisor) in 2005 and took on an SMA (Separately Managed Account). So, we’ve got two businesses now: one is stock market research business and the other is traditional asset management business.
Do you have any competitors?
JR: I hold a patent on our computational engine for simultaneous analysis by multiple guru investment strategies. All the industry will say that this is the model they like. However, most companies promulgate just one model — I took the broader approach of including many successful strategies.
What about future guru additions to your engine?
JR: We’re very slow to add gurus. We’ll test new gurus internally for a number of years.
What your take on expert investment communities like the Fool’s CAPS or Covestor?
JR: I’ve found that there is no predictability between the top 20% of performers in one period, the hot hands, to what happens in the next period. I think investors need long track records of 10+ years — there is going to be very little of a success factor when people follow anything less.
What about market neutral strategies, meaning investors who short stocks as well as buy them long?
JR:I haven’t found one guru who’s written about shorting with an established track record. I also don’t address endowment managers like Yale’s Swensen because his strategy employs things outside of equities, like Real Estate and Private Equity, which are all areas I don’t cover.
After following multiple strategies over the years, do you have any interesting takeaways?
JR: I’d like to share a couple of interesting things we’ve learned.
First, there are several approaches successful over the long term. To my surprise, though, there are many successful models that exist simultaneously. But, you have to be disciplined to stay the course of a particular strategy over the long run, even when severely hurt by the market in the short term. Many investors don’t stay the course. For these strategies to beat the market, investors have to stick with them. David Dreman and Ken Fisher all say that you have to ride it out. No matter what, we’ve found that you have to resist modifying the strategy.
Secondly, every single guru has had a down year or two in a row compared to their benchmarks. Nevertheless, these gurus beat most indices and other investment professionals over the long term. If you read Greenblatt, he makes it very clear that investors will encounter 1 or 2 years which appear not to be working. Given the fact that the systems always seem to snap back, I also wondered myself what success investors would have in investing in gurus after a smashing year. What if you invested after a year of losses? I found that in most cases, guru strategies returned either in in that year or caught up next year caught up (made up for losses).
How is your general performance [ed. this question was posed in September 2008]?
JR: At this moment, our 2 top strategies with highest annualized returns 5+ years are Benjamin Graham and Ken Fisher’s SuperStock (p/s) up over 21%. while the S&P is up just over 3%. It’s a huge difference. In fact, all strategies have outperformed the S&P 500.
How about tumultuous years like this one?
JR: Graham and Fisher continue to be best performers. That said, I found no price predictability in short periods (3 months). If you had looked at these at different points of time, you would have found down years throughout the period.
One of the things I want to emphasize is that Validea’s guru models implement a lot of AI that is not present in Screening 1.0 products. I’ll give you an example: one of William O’Neil’s rules is that earnings must increase year over year unless there is a dip one year and earnings recover to new highs. That kind of intelligence is part of the O’Neill guru model. Another example: determining an earnings growth rate is not simple. Let’s assume we drill down on a 3 years earnings growth rate. What if you begin with 0? Our system adjusts to be able to assess the stock and make it useable the same way as human analysts deal with data. There is a major shortcoming in Screening 1.0 — databases are not clean, don’t have all data points, and plus, the databases can’t adjust for certain criteria which makes it hard to stay on scale.
Thanks for participating, John.
