Methodology for Dealing with Ambiguous Advice
Methodology for Dealing with Ambiguous Advice
By definition, when a newsletter’s advice is ambiguous, it is impossible to ascertain every last detail of the portfolios the advisers want their subscribers to construct. To deal with each case of ambiguity impartially (since we have no stake in one or another letter doing well), the following rules were adopted years ago by the Hulbert Financial Digest. They set down what we would do in the event a newsletter was silent or vague about this or that aspect of constructing a model portfolio.
You should realize, however, that there is no one right way for a subscriber to deal with vague and ambiguous investment advice. One of the inevitable consequences of ambiguity is that different subscribers, each faithfully following such advice, nevertheless may invest their portfolios in quite different ways—with accordingly different results. To the extent you would have interpreted a newsletter’s advice differently than what is reflected in these rules, your results would vary from those published here.
Our rules for dealing with ambiguous or incomplete advice
Most of our rules come into play only in the event a newsletter’s advice in some way is silent or vague. If it clearly and unambiguously deals with all aspects of translating its advice into a portfolio, then we follow that advice.
In short, we construct portfolios for non-model-portfolio letters that have the following characteristics, unless the newsletter advises specifically to the contrary:
- is fully invested;
- employs no margin;
- gives equal weight to each position;
- includes just those securities most highly recommended at any given time.
If an editor wants to have a certain percentage of subscribers’ portfolios out of the market and in cash, wants the model portfolio to be margined, or wants unequal allocation of the portfolio between its various components, and so on, then we require the editor to say so specifically.
Implicit in this approach is that we take a “total portfolio” approach to rating newsletters. In other words, we believe that the best way to measure a newsletter’s performance is by trying to decide how much to keep in cash, what weight to give to each position, and so on. To put it another way, it is not enough to say “the average recommendation of newsletter ABC gained X%.” Such a statement does not take into account the importance of a security’s weight in a portfolio and the order in which securities are bought or sold in the portfolio.
Also implicit in our rules listed is that the hypothetical portfolios we construct undertake certain rebalancing transactions that no one subscriber to a newsletter is likely to undertake.
To understand the need for these additional transactions, consider a service which recommends purchasing a new stock without also selling a currently-held position. Where are you to get the money to buy the new stock? Undoubtedly, each subscriber will deal with this question differently, some selling out this or that security, some selling out partial positions in several securities, and some deploying new amounts of cash not previously invested according to the advice of this particular newsletter. Taking these various subscriber responses into account, what portfolio weight will the new recommendation have relative to other securities in the portfolio?
The least arbitrary response to this question is to assume that the new recommendation will have the same weight in the portfolio as the other securities in that portfolio (unless the newsletter specifically advises to the contrary). Therefore, after buying the new recommendation, the HFD undertakes a number of rebalancing transactions so that thereafter all securities in the portfolio enjoy equal weight. This means that if a stock has gained enough in value to have greater-than-equal weight, a portion of it is sold to bring it back into line with the others. And if a stock has declined in value so that it has less-than-equal weight, more shares are purchased to bring it back to the same weight as others. (Commissions are not charged on rebalancing transactions).
New vs. Old Subscribers
One way of thinking about the HFD’s response to the above issue of rebalancing is to take the perspective of a new subscriber to the newsletter in question. Confronted with a list of recommended securities, the new subscriber intent on following the newsletter’s advice will divide his or her assets equally among them—regardless of whether one of those recommendations is a newly-rated “buy” that was unaccompanied by a “sell.” The portfolio constructed by the HFD for this letter, after undertaking the various rebalancing transactions, will look just like the one constructed by the new subscriber. In general, other things being equal, the portfolios constructed by the HFD for non-model-portfolio letters will look like the ones constructed by new subscribers.
The different perspectives of the new and old subscriber also come into focus in the HFD’s treatment of “buy” and “hold” ratings. As pointed out above, the HFD constructs portfolios for non-model-portfolio letters out of those securities most highly rated by them. This means that if 50 stocks are on a recommended list, and of them 25 are rated “buy” and 25 are rated “hold,” the portfolio the HFD constructs will include just the 25 “buys.” And when a stock is downgraded from a “buy” to a “hold,” the HFD’s hypothetical portfolio will sell that stock. (The only exceptions to this rule come if and when a newsletter specifically says that their stocks rated “hold” are just as highly recommended as their stocks rated “buy,” in which cases the HFD’s portfolios include both the “buys” and the “holds.”)
The HFD’s orientation towards the new-subscriber perspective helps to explain this treatment of “hold”-rated securities. A new subscriber presumably will buy just those securities rated “buy,” while longer-term subscribers may or may not own the securities rated “hold”—depending upon the length of time they have been subscribers and if they were following the newsletter’s advice at the time those securities were rated “buy.” How long should the HFD carry a “hold”-rated stock in a portfolio that is supposed to be representative of a wide variety of subscribers? Should we automatically sell it if it hasn’t been mentioned as a “buy” within, say, the last three months? Or should we wait six months or a year? Rather than legislate an arbitrary cutoff for selling “hold”-rated securities that have been held for a certain period without an intervening “buy” recommendation, the HFD instead takes the perspective of the new subscriber and constructs the portfolios out of just the “buy”-rated securities.
To illustrate the pitfalls that would await the HFD if it were not to treat “hold” in this way, consider a letter that on January 1, 1980, recommended IBM as a “buy.” Assume further that it downgraded IBM to a “hold” on February 1, 1980, and has carried it as a “hold” in every issue up to the present. How many subscribers to that letter would have this stock in their portfolios today? If they faithfully followed the letter’s advice, only those who were subscribers during January 1980 would own it. For all other subscribers, the performance of IBM is irrelevant.
Replies to the Skeptics
Over the years the HFD has heard a number of objections to its treatment of “hold”-rated securities, and you should be aware of the HFD’s response. In that way you will know how to respond if and when particular newsletters use these objections as a way of dismissing their HFD performance ratings.
One criticism of the HFD’s treatment of “holds” is, simply, that “hold” means hold and not sell. But the point to bear in mind is that “hold” is ambiguous, and that there is no way of avoiding treating “hold” as something other than hold. To say that “hold” means hold misses the point, which is that hold’s meaning itself is not clear.
Consider, for example, the consequences to new subscribers of deciding, in contrast to the HFD, that a “hold” should be treated as a “buy.” It would entail having them buy all of a letter’s recommendations, the “holds” as well as the “buys.” But if “hold” doesn’t mean sell, then why should it mean buy? Is a “hold” a “buy” or a “sell?” It can’t be both. The HFD’s critics on this issue have resolved nothing with their suggested “solution.”
Another objection to the HFD’s treatment of “holds” is that it causes Performance Ratings for newsletters to be lower than otherwise. Howard Ruff of The Ruff Times articulates this objection as follows: “If I recommended Squibb as a buy at 64, and made it a hold at 70 until it reached 90, Hulbert’s program would close me out at 70.” But Ruff’s criticism fails to focus on what the HFD does when, to use his example, Squibb is downgraded to a “hold” and is sold. The proceeds are not stuffed into a mattress, thus preventing Ruff from making more profits. Instead, the HFD reinvests the proceeds into stocks Ruff is rating more highly at that time. This is crucial to understand. The HFD’s approach has the effect of keeping Ruff’s portfolio invested in nothing other than the securities he’s most highly rating.
An example from Ruff’s own use of “buy” and “hold” illustrates this important point. In November 1988, after several months of highly recommending several Australian investments, Ruff’s enthusiasm lessened. Explaining that “The Aussie dollar could slide if the greenback rallies” and conceding that he wasn’t “convinced the U.S. dollar bull market is over,” Ruff downgraded his recommendation on his Australian investments from “buy” to “hold.”
Consider, thus, what happened to Ruff’s portfolio when the HFD reoriented it out of these Australian investments into the other securities that Ruff liked better at that time. It was putting Ruff’s best foot forward. The only way that the HFD’s treatment of his portfolio could cause his performance to be worse would be for his Australian investments to perform better while rated “hold” than the securities he rated “buy.” But if that is the case, Ruff shouldn’t have downgraded them to a “hold” in the first place. The finger of blame should not be pointed at the HFD in such a case. It was he, not the HFD, that believed that these Australian investments had less potential and therefore should be downgraded to a “hold.”
What is true for Ruff is true for all the non-model-portfolio newsletters. In general, the HFD’s approach to “buy” and “hold” would reduce a newsletter’s performance only in the event its recommendations performed better while rated “hold” than while rated “buy.” But why should that be the case? It is not the HFD, but the newsletters themselves, that have chosen to downgrade stocks to “hold.” It is their editors who have decided that other stocks, rated “buy,” are better bets than those downgraded to “hold.” The HFD is taking them at their word.
Further Replies to Skeptics
Another criticism the HFD has received over the years concerns the issue of rebalancing. (As discussed above, for non-model-portfolio letters these rebalancing transactions entail selling off small portions of a portfolio’s better performers and buying more of a portfolio’s poorer performers.) Most commonly, editors have articulated their objection to this practice by arguing that it violates the cardinal rule to “let your profits run.” Of course, there is another investment cliché that runs: “Buy low, sell high”—and which is directly contradictory to these editors’ cardinal rule. But the point is not which cliché is best; rather the point is that rebalancing is necessitated by ambiguous advice. If editors want to let their subscribers’ profits run, all they have to do is say so.
Consider the case of a letter that, just prior to the 1987 crash, was recommending that 5% be invested in out-of-the-money puts. Because of the Crash, of course, those puts skyrocketed; by mid-November, in fact, when this letter’s next issue appeared, they represented nearly 50% of the portfolio’s value. What was a subscriber to do upon reading in that new issue, “continue to have 5% of your portfolio invested in put options”? Should he sell off the bulk of those put options to bring their percentage weight back down to 5%? Or does the adviser want the subscriber to continue to hold what was originally 5% but is now 50%? A case could be made for either course of action. Ambiguity has struck again.
Some letter editors, recognizing this ambiguity, have clarified their advice. They’ll tell their subscribers, for example, that their puts which originally represented 1% of their portfolio now represent 5% of the portfolio, or whatever. If that is too heavy an investment in those puts, these advisers then would sell off a portion of them. Similar options for clarifying their advice are open to any adviser, of course. All they must do is say in their newsletters what they mean—and not keep the HFD (or their other subscribers) guessing.