Analyzing the Analysts - Part II

Includes: SPY
by: Braun Research

<< Return to Part I

The users of the system described in Part I often apply other methods than those used in managing CAP to qualify the candidates detected for them for further analysis. These include - but are not limited to - DDM approaches, other proprietary or commercially available DCF models, technical analysis and who knows what else each analyst or PM comes up with?

DDM stands for Dividend Discount Model while DCF stands for Discounted Cash Flow. They are standard formulas used extensively by professional analysts to calculate the current "fair" value of a stock by using either expected dividends being paid (DCF) or expected cash flows (DCF) as provided by various services or estimated by the analyst.

Quite often, no more than good-old-fashioned analysis is applied to the limited and therefore manageable universe pre-selected as described in Part I. The benefit is to maximize the productivity of the analysts who represent everywhere a scarce and expensive resource. It is not a very good idea to pry into what others do with the data or input, as no one in this industry takes too kindly to such curiosity. Some users get better results than the CAP proof-of-concept portfolio did and that's the way it should be.

The selling disciplines are not addressed here because professional managers tend to have their own sell disciplines. Just for completeness, let it be mentioned that the CAP turnover is about 50% annually, meaning that a typical stock is expected to be held about two years. Selling is guided by a very disciplined risk control process, which may boot out stocks earlier. Also stocks reaching their full estimated intrinsic value are sold.

The full understanding of the system does require a two-three months period of getting used to its common-sense but somewhat unconventional way of thinking. The process uses a pattern-matching technology and is defined as a "binning" approach, as opposed to an "exposure," which is what a low P/E manager would use for example or what a Barra analysis and the majority of screening methods provide. Nothing wrong with those, except that too many managers tend to use them and they end up therefore with a tendency to be correlated; in other words part of a herd: not the best thing when attempting to beat the market.

Before wrapping up, I would be remiss not to mention a further interesting use of the system. The CAP proof-of-concept portfolio does make implicit sector allocation decisions. These bottom-up allocations turned out to be remarkably prescient. At the top of the market in the 90s, the system took for example users out of tech and put them into consumer stocks. This was highly contrarian as major wire-houses were telling their clients at the time that they will suffer badly because consumers will be kept away from the malls by the fear of being "Anthraxed." Remember the Anthrax hysteria? No such thing happened! And we said so in advance. In 2006-2007 the allocation was screaming: CASH! This was when the markets kept going up and the pundits did not see any danger on the horizon. 2008 however was lurking around the corner. Of course, professional investors can rarely go to 100% cash; neither did the CAP. But, through attrition, it went to 30% cash, simply because the system was not finding anything worth buying while selling was being recommended by the discipline. Those who elected to listen, lowered their portfolios' betas as much as they could. In 2009 the portfolio's implicit sector allocation caught (almost) the bottom. As for the past year or so, manufacturing stocks were bought at double their S&P500 weighting for some rather decent returns. Now the discipline of the process "rotates" the portfolio out of industrials as illustrated in these two successive bar charts:

Click to enlarge

There are certain users of the system, typically in very large organizations, who, through their function as strategists, do not care too much about the individual stock selections made but are faithful followers simply for this sector allocation feature.

This however is not your usual sector allocation approach. Normally strategists proceed in a top-down manner, guided generally by government reports. Thus, by the very nature of such reports, the information conveyed to the investors is delayed anywhere from one to three months as compared with CAP's bottom-up process, which is much more timely and, through its equal weighting, is very sensitive to the economy as a whole rather than to the largest-of-the-large companies, which dominate and drive most indexes used by most strategists.

This is not to say that top-down analyses are not valuable. All I am trying to convey is the idea that this rather unique bottom-up approach is a valuable complement to the top-down analyses: when there is a coincidence between the two, you know that you have a much stronger signal than considering either, separately.

It is interesting that the manager of CAP has been itching for some time now to trim the manufacturing holdings as they appeared to become somewhat "long-in-the-tooth." But the discipline did not permit it. And, by long experience, he learned that the system was more intelligent than we humans are. The system was proven correct once again: selling when the manager thought was appropriate, would have been too early!

When will the next top occur? Hard to say. But it is not likely to be very soon (5-7 years?) and it is likely to be forecast by this bottom-up approach, in time for effective defensive action. I wrote recently another article on Seeking Alpha, dealing with this subject. A more detailed performance graph of the portfolio was also shown there and I include here an update:

Click to enlarge

In conclusion, the idea of analyzing the analysts, as pursued and developed by its originator for more than 25 years, has proven to be a worthwhile endeavor and the CAP's performance confirms Professor Bradshaw's intuition.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.