An animated debate on geometric vs. arithmetic attribution

Here's a new way to communicate ideas, issues, and controversial topics. Hope you like it!

Riskless stocks

In this weekend's WSJ, Jason Zweig reminds his readers of James K Glassman's 1999 best seller, Dow 36,000: The New Strategy for Profiting From the Coming Rise in the Stock Market (which, by the way can apparently be had at Amazon for a mere one penny!). Jason mentions that in his new book (which I won't bother to reference), he admits that he "was wrong."

Recall that Glassman picked up on the hot bull market of that time and made this bold, and apparently rather unrealistic, prediction for the future of the DJIA. Glassman and his coauthor, Kevin A. Hassett felt that "stocks and bonds are equally risky over long horizons," because "stocks had never  lost money over the long term." Zweig further points out that investors back then had apparently learned that "socks weren't risky."

Zweig cites LA-based Oaktree Capital Management's chairman, Howard Marks, who said that this logic is false. Zweig also alerted us of a new "superb forthcoming book" by Marks that "helps explain risk clearly": The Most Important Thing. The book is due out May 1, and I preordered my copy today.

Rather than repeat here any further commentary from Jason, I will leave it to the reader to link across. Some insightful points are raised which are worth having a look at.

A bit of history on the NYSE

When I was an undergrad there were only two subjects I considering majoring in: history and mathematics; I chose the latter, but have continued to enjoy the former. In today's Wall Street Journal, Jason Zweig begins with a reference to the possible NYSE Euronext and Deutsche Börse AG merger, but quickly moves on to provide us with a bit of history on the NYSE and the role of exchanges.

Jason is an excellent writer, and I very much enjoy his columns, and today's is no exception. In a fairly crisp manner he discusses some background on the exchange and how the they have evolved. It's definitely worth your while to have a look.

More support for changing the way we do things ...

I love the weekend edition of The Wall Street Journal. Invariably I find one or more articles to reference in one place or another. And this past weekend it was Matt Ridley's "A Key Lesson of Adulthood: The Need to Unlearn." As Matt so correctly put it, "We all think that we know certain things to be true beyond doubt, but these things often turn out to be false and, until we unlearn them, they get in the way of new understanding." It's ironic that this article appeared as President George W. Bush's former Defense Secretary, Donald Rumsfeld's Known and Unknown: A Memoir was being released. Rumsfeld spoke of the "unknown unknowns," as things we don't know that we don't know. While he was ridiculed for what some thought was a silly statement, it has a great deal of value.

In our slice of the investment universe, many performance measurement professionals should strive to be open to unlearning concepts they've embraced for years or, in some cases, decades. I have commented at length in this blog, our newsletter, and articles about the undeserved adulation awarded to time-weighting, and how money-weighting should be the most commonly used method to derive returns.

Matt Ridley cites the word "disenthrall" from Mark Stevenson's An Optimist's Tour of the Future," who borrowed it from an 1862 message from Abraham Lincoln to Congress, suggesting that they disenthrall themselves from "the dogmas of the quiet past," meaning to "think anew." Well, performance measurers should disenthrall themselves from the belief that there is one way and one way alone to measure performance and think anew! And, for that matter, that GIPS(R) (Global Investment  Performance Standards) hold the key to everything we do in performance measurement. We very much need to "think anew" about much of what we do.

In defense of quarterly verifications

A recent blog post questioned the value of having quarterly verifications done. Interestingly, we received the following response: "Running the verification on a quarterly schedule can help catch any possible errors sooner, rather than later, potentially minimizing the impact of that error and preventing it from compounding as the firm awaits for the annual verification review." I found this interesting for a few reasons.

I cannot deny that this could be a benefit; however, I question why errors would be appearing that require the verifier to uncover. The firm should have controls, policies, and procedures in place to ensure the accuracy of their returns. In addition, the annual verification provides confidence that they are following the rules of the Global Investment Performance Standards (GIPS(R)). But other thoughts occurred to me, as well, which I wish to share.

First, given that we have heard that one verifier who promotes quarterly is several quarters behind for many of their clients, if a client is hoping to learn of errors in a timely manner, it appears this might not be possible.

Second, verification tests whether the firm's processes and procedures are designed to calculate and present performance results in compliance with the standards. It would seem that the annual review should affirm that this is the case; what would cause the firm one or two quarters later to suddenly diverge from their processes and procedures, resulting in errors that require the verifier to discover?

Third, many of our clients previously had quarterly verifications done; they no longer do, suggesting that they prefer annual (because we will do quarterly verifications if they want).

No doubt events can occur during the year which require the firm to have to make decisions which might result in problems, such as the discovery of errors, the creation of new products, acquisitions, or simply questions that arise from within the firm. I would expect that the firm's verifier would be available to respond to questions and offer guidance; we do. We often get calls and emails throughout the year from our clients, and are happy to offer our opinions and recommendations.

Clearly, if the firm desires to have quarterly done, then the verifier should be happy to accommodate them; we would. We just happen to feel that such frequent visits are disruptive, more costly, and accomplish little, if anything. Yes, it keeps the verification firm's staff busy throughout the year, but beyond that, where's the value? We firmly believe that annual is sufficient and the ideal timing for verification.

Is GIPS at the center of the investment performance measurement universe?

At times it appears that many individuals in performance measurement, as well as those on the periphery (e.g., portfolio managers), see GIPS(R) (Global Investment Performance Standards) as being at the center of everything we do. It's quite common to hear, for example:

• is our client reporting compliant with GIPS?
• is our custodian compliant with GIPS?
• are our returns compliant with GIPS?

"GIPS, GIPS, GIPS." That's the chant or mantra.We've experienced brokerage reps asking whether their firm complies with GIPS, even when the question doesn't fit. Custodians and GIPS? Hardly "perfect together."

And while like Copernicus and Galileo I am risking a charge of heresy or blasphemy to suggest otherwise, I feel compelled to do so: no, the world of performance measurement doesn't revolve around GIPS; there, I've said it! But at times it definitely appears to be that way. But just as the geocentric or Ptolemaic views were based on what appeared to seem quite intuitive and logical, the same kind of erroneous view is wrong regarding GIPS, in spite of similar apparently obvious appearances or what your intuition suggests.

GIPS has become the sole performance measurement standard that is referenced these days, in spite of the fact that others (e.g., the Bank Administration Institute's from 1968 and Investment Counsel Association of America's (now the Investment Adviser Association) from 1971) have existed and not been repealed. Clearly there is room for more guidance, but GIPS isn't the "end all" for all things performance. And while it serves as a useful tool to gauge how firms are doing, it's not always the ideal. For example, in the world of brokerage, where most of the firm's clients are non-discretionary, to provide time-weighted returns is ill advised, contrary to the GIPS standards (which speaks not to this part of our universe).

GIPS is an extremely valuable standard which we obviously support and often comment on. We have loads of clients who comply and encourage others to do so as well. However, to think that GIPS is a panacea that answers all questions related to performance is simply incorrect. Unfortunately, like the old saying that if the only tool you have is a hammer, all problems look like nails, it appears that GIPS is the sole standard and therefore the answer to everyone's performance question; but it isn't.

Two (or more) sets of books

I conducted a GIPS(R) (Global Investment Performance Standards) verification for a client on Monday and was asked if it made sense to calculate returns in multiple ways. That is, if they know that the client's custodian or consultant calculates returns using a different method than prescribed by GIPS, would it be prudent to maintain a separate return stream which they could turn to in the event they need to reconcile?

I was intrigued by this question as it hadn't previously been posed, and I am unaware of anyone who does this.

There is absolutely nothing wrong with maintaining multiple sets of returns. But, I recommended that they don't do this as a matter of standard practice, but rather only when approached to justify their return or reconcile to one (or both) the other parties. Why do all this work when you don't know if it will be needed? Wait and then do the appropriate calculations and analysis.

There are many reasons for potential differences in returns, and the return formula is just one. Many asset managers are often charged with doing this sort of thing and perhaps if they know that a particular client religiously asks them to prove out their numbers, then in that case perhaps doing it on an ongoing basis makes sense, but otherwise, I'd do it ad hoc.