Is GIPS just for the institutional market?

I learned that some folks feel that there's a perception in the industry that GIPS(R) (Global Investment Performance Standards) only applies to institutional managers, and that retail need not apply. I found this somewhat fascinating, bewildering, surprising, and clearly erroneous.

In the early-to-mid 1990s, a firm that marketed to institutional clients had an ADVANTAGE by claiming compliance with the standards (the AIMR-PPS(R) back then), but today they're AVOIDING a DISADVANTAGE by complying.

However, in the retail space a manager has an ADVANTAGE by complying because most firms catering to this market don't. But this isn't because the standards don't apply. If your clients are high net worth individuals, clearly most (if not all) will never have heard about GIPS, which is fine. YOU can educate them about the standards. That they're a set of ethical principles that foster full disclosure of information; that they're globally recognized; that you invested significant amounts of time and money to comply; that they are viewed as the "best practices" for reporting performance to prospects; that you want to comply because you feel that by complying you're providing your clients and prospects with the best information to gauge your past performance. I could go on-and-on, but hopefully you get my point.

GIPS isn't just for institutions any more...the retail market is huge and waiting for managers to hop on the GIPS bandwagon.

"Classics" reviewed

Classics in Investment Performance Measurement (The Spaulding Series), the book that Jim Tzitzouris and I edited, has been reviewed by Jerry Tempelman, CFA. Jerry's review appears in the current issue of the Financial Analysts Journal.

We are obviously quite pleased that our book got this attention. The idea for the book came to both Jim and I individually, and we decided to collaborate on its creation. It runs close to 400 pages and includes articles on performance measurement, attribution, and risk, that span several decades. We suggest you read Jerry's review to get an independent assessment  of the book.

p.s., Want a copy? Contact my office (732-873-5700) or e-mail Patrick Fowler and tell him you're a reader of this blog, and we'll knock $30 off the purchase price!

GIPS Week in San Francisco!

This week sees the GIPS(R) (Global Investment Performance Standards) Executive Committee (EC) meeting in one of my favorite cities: San Francisco. In addition, the USIPC (United States Investment Performance Committee) will hold its annual "in-person" meeting, and the Americas' RIPS (Regional Investment Performance Subcommittee) will meet. AND, there's the annual GIPS conference, an annual event that many look forward to and which our firm is sponsoring, being held here, too. And so, lots of  "GIPS stuff" is happening this week. I'll share with you some of the insights I obtain over the next several days.

When's a composite not a GIPS composite?

When it's an "administrative composite."

I just commented on a response to someone who commented on yesterday's post, regarding "catch-all" composites. Many firms utilize "administrative" composites to keep track of accounts that aren't in GIPS(R) (Global Investment Performance Standards) composites. For example:

• accounts below the firm's minimum
• non-discretionary (for GIPS purposes) accounts
• accounts transitioning from one composite to another
• accounts that haven't yet been added to a composite
• accounts that have terminated.

Administrative composites can be helpful tools to manage your compliance. However, these composites aren't considered "GIPS composites," meaning that you

• don't prepare presentations for them
• have them listed in your "list of composites."

Admin composites aren't required, though they can be helpful. 

p.s., if an account that should be in a GIPS composite isn't, but is in an Admin composite instead, the requirement that it be in a composite won't be fulfilled.

p.p.s., "catch-all" composites aren't permitted. However, if the firm uses these to bring together all those accounts which are non-discretionary, that's fine.

Catch-all composites are alive and well...

In 1993, while attending a conference on the AIMR-PPS(R) standards, a speaker made the statement that virtually all compliant firms would have "turkey composites." I was a bit dumbfounded by this, because at the time I didn't know what a "turkey composite" was. Well, I came to learn that it meant a "catch-all" composite: that is, a composite where firms could stash all those accounts that just didn't seem to fit anywhere else. This concept made no sense to me, since composites are supposed to comprise accounts with similar investment styles; not be a place for misfits. Eventually, AIMR announced that such composites weren't permitted.

That's all well and good, but the reality is that there are still firms that use them, even though GIPS(R) (Global Investment Performance Standards) also disallows them. This week we learned of an asset manager that has such a composite, even though they're verified! But we know that this particular verifier has a "soft spot" for such things, and apparently doesn't mind if their clients want to avail themselves of them. Interestingly, this asset manager actually knows themselves that they're not permitted, but still has one.(The SEC, of course, may have a different opinion than this firm's verifier, but that is an entirely different matter).

But what does this actually mean? Well, in a nutshell it means that the firm isn't actually compliant, because having such a composite invalidates their claim.

Yes, these composites are a nice tool to put those accounts that just don't seem to fit anywhere else; and yes, they save on having lots of small composites, but unfortunately, they're not allowed. But clearly that isn't stopping firms from still using them.

A place for TWRR and subportfolio returns

If you're a frequent reader of my blog and/or our firm's newsletter, you know of my passion for money-weighting, especially when it comes to subportfolio returns.

I'm teaching classes this week for a client and when we discussed subportfolio returns and my preference for money-weighting, a student asked what the GIPS(R) (Global Investment Performance Standards) rules were regarding this topic. In general, GIPS is silent when it comes to subportfolio returns; the one exception is for carve-outs. But even here, there is no explicit requirement, but it's worth considering how we should handle them.

While I generally favor using MWRR for subportfolio returns, in this (possibly only) case, I'd argue that TWRR is what we should do. And why is this? Well, think about what a carve-out is supposed to be doing: we're taking, for example, the equity portion of a balanced portfolio and putting it into an equity composite. If there are accounts in the composite which are only invested in equities, their returns are measured using TWRR. And, the carved-out portion is supposed to represent what would have happened had the assets been managed separately, and not as part of a balanced portfolio. Therefore, TWRR makes sense.

Negative Sharpe Ratios

A problem surfaced over the past couple years that has long been lingering within the world of performance measurement and risk: the potential for "negative Sharpe ratios." While I had heard of a problem, I hadn't seen it first hand until about two years ago and then fully understood the difficulty.

When both the portfolio and benchmark have negative returns, we can have negative Sharpe ratios and the results can be counter intuitive. I wrote about this in our newsletter two years ago and have been discussing it ever since, especially in our Fundamentals of Investment Performance class.

Well, this week I'm teaching the class for a client, and we touched on this today. I have rationalized WHY negative Sharpe ratios make sense and will shortly be writing at length on this topic. Consider this more of a "teaser" as to what to see in the near future.

How's your index data? Is it clean enough?

I've been having a discussion with a colleague on the topic of index data. I suspect that many asset managers don't give these statistics much thought, simply taking in whatever values they're given by their data providers. But should they?

Many performance measurement professionals recognize that there can be problems with index data. Some managers track the constituents themselves and identify problems when they occur. In some cases they reach out to the index provider and alert them. But do the vendors correct the problems? Sometimes, yes; but not always.

Given that indexes play an important role in performance measurement (for example, we use them in attribution, in client reporting, and in our GIPS(R) (Global Investment Performance Standards) presentations), should we be paying them more attention? Or, do we think that (a) the errors are immaterial or (b) that over time they'll work themselves out?

Any thoughts?

"The strange position of risk manager"

I have become used to some of Nassim Taleb's seemingly outlandish remarks, and so was not terribly surprised when reading Fooled by Randomness: The Hidden Role of Chance in Life and in the Markets that he would make such a statement. To try to at least start to bring this statement into context, I should provide a bit more of the section from where it was drawn:

"Corporations and financial institutions have recently created the strange position of risk manger, someone who is supposed to monitor the institution and verify that it is not too deeply involved in the business of playing Russian roulette. Clearly, having been burned a few times, the incentive is there to have someone take a look at the generator, the roulette that produces the profits and losses...The risk managers' job feels strange: As we said, the generator of reality is not observable." (page 40)

I'll leave you the task of pursuing the full and complete context if you so desire. But Taleb does raise (as he often does) an interesting question regarding the appropriateness and value of a risk manager, as well as what his/her duties should actually be and how they might carry them out.

Can risk be managed?, is essentially the question he is asking. And so, can it? Should it? How to do it? If you agree with Taleb, I'd love to hear from you.

The floor is yours...

Transitioning from one composite to another

I conducted a GIPS(R) (Global Investment Performance Standards) verification for a new client earlier this year, who had previously used a different verifier. During the course of the review, I discovered a number of items that prohibited me from issuing a "verification report," including their policy on switching accounts from one composite to another, when a client's objectives had changed. They would keep the account in the old composite throughout the transition period, and move it out at month-end, when it had fully transitioned, and on that same date move it into the new composite. I explained that once they began the transition, they had to remove the account from the old composite. They requested a more formal justification for the basis for my position, and I will deliver it to them shortly. But I thought it worth discussing here, as well, as this may be a problem others are dealing with.

The client provided me with an example which shows the account's transition:

In this example we can see that on or about the 20th the shift from the old strategy to the new began.

This matter goes to the whole subject of composite definition and makeup. A composite consists of accounts with similar investment styles, strategies, etc. The members of the composite take on the identity of the composite’s definition, thus qualifying their presence within the composite. The moment you take an account and begin to transition its identity to that of another composite is the moment the account begins to no longer look like the composite from where it came (i.e., to lose the earlier identity). The transactions the firm is making to alter the account’s identity are ones that they would otherwise not be making; the identity of the account would have remained intact, had the client not requested a change in style. But in honoring this request, they are altering the account’s identity such that during the transition period, it is losing the identity of the original composite and taking on the identity of the new one, so that when they are complete it now fits into the new composite.

And yet they still have that account, along with the sales that occurred, in the original composite, even though it lost that account’s identity; consequently, its return uses transactions and an ending market value that don’t belong. On the final day it doesn’t resemble the other accounts in the original composite; it doesn’t have the identity of the composite and clearly doesn’t belong. Consequently, this account does not belong in the original composite on the end date nor during the transitional period.

Just as with a new account, you don’t bring it into the composite while you’re transforming it from its original state to the composite identity, you don’t move the transitioning account in until it has taken on the identity of the composite it is moving to (as you correctly do). And, just as with an account that is terminating, where you don’t keep it in the composite while you are liquidating its assets to meet the client’s request, taking actions you wouldn’t otherwise do, and keeping the account in the composite, the account that is losing the identity of the original composite cannot stay in that composite while you are transforming it to the new identity

On page 101 of the 2006 edition of the GIPS Handbook we find “Firms are not permitted to include portfolios with different investment strategies or objectives in the same composite.” At the end of the month the account has transitioned to the new composite’s strategy, and yet is still in the old composite. This is a violation of this prohibition.

There's also a Q&A which addresses this topic:

One of the portfolios in our All Cap Growth Equity composite changed objectives towards the end of the second quarter (client agreement date: June 21) from All cap ($300MM and above market cap) to Mid Cap ($1B - 10B).

Should the portfolio be taken out of the All Cap composite for June, even though the portfolio’s objective only changed in the last 9 days of the month? In talking to the portfolio managers involved, no changes were implemented during the short period because of the similarity of the two styles (all cap vs. mid cap)

Does the client objective change automatically exclude the account from being in the all cap composite for June?

Shifts from one composite to another should be consistent with the guidelines set forth by the specific account agreement or with documented guidelines of the firm's client. In the situation you present, if there was any latitude within the agreement as to when the change should occur, in the interest of fair representation and full disclosure, the firm should include the account in the All Cap Growth Equity composite only through the last full measurement period that the account was managed to that strategy.

Do the standards explicitly address the issue of timing when it comes to composite switching? No. However, there is clearly enough supporting material to justify the requirement that once an account begins to transition from one strategy to another, it needs to be pulled.

You might ask, "well, I thought that all actual, fee-paying, discretionary accounts have to be in at least one composite; if we remove the account during the transition period we're violating this rule." You're correct about the rule, but there are exceptions and this is one of them.

Sensitized to the lowest common denominator

Earlier this year I attended a football (soccer, for you Americans) game at Wembley Stadium in London. During half time my host informed me that we weren't permitted to take our beers into the "pitch" (to our seats, in essence) because of prior problems with rowdy fans. In other words, all fans suffered because of the actions of a few bad folks.

This reminded me of an incident that occurred while I was in the Field Artillery more than 30 years ago: we had been out on a training mission and were wrapping up, after what had been an excellent day. Everyone was upbeat and excited to be heading back. It was determined that some powder bags were missing. The XO (Executive Officer) decided to line everyone up at attention until either the guilty party, or someone who knew who the guilty party was, stepped forward and admitted his guilt. I was merely the AXO (Assistant Executive Officer) and although I disagreed with this action, could only offer my thoughts and nothing more. After about an hour of this silliness, the decision was made to head back, even though (as I recall) we didn't know who had committed this violation. What I do recall was that morale sank as a result: again, the majority who were innocent suffered because of the actions of a few.

Sometimes standards are developed with great sensitivity to the worst of our lot rather than the average or above average individual or firm. This is unfortunate. When inconvenience and added costs are piled on because we have to be concerned with those who don't abide by the rules or who attempt to take advantage of others. What a pity. [Please don't construe this post as a criticism of GIPS(R) or any other standard; it is simply a broad statement].

What is an INSTITUTIONAL account?

I just got off the phone with a GIPS(R) (Global Investment Performance Standards) verification client, who is splitting away their institutional business from the trust side of their firm: a common step for banks and trust companies. They must decide on what an institutional account is.

First, there is no single, uniform definition: each firm must decide on what this means for them. They must document the definition and ensure that it's applied consistently.

What might some of the criteria be for the definition? I suggest:

1. The type of account: for example, endowments, foundations, and pension funds are typically thought of as "institutions." Clearly some "high net worth" individuals can qualify as institutions, too.
2. The size of the account: firms might establish an initial threshold to use to split away the "institutional" from the non-institutional. For example, perhaps $1 million or $5 million.
3. The source of the business: many firms market their services through third-parties, who may offer their clients wrap fee or UMA (uniform managed account) products. In these cases, the third party provider can be viewed as an "institution," even though the end client may be retail. These providers are, in a sense, a client, too, and often require managers to comply with GIPS.

Banks usually have "trust agreements" which make the trust accounts fairly easy to identify, though there may be accounts that still need to be split away. Defining criteria is an important and necessary step in this process of making sure the institutional accounts are properly segregated into the "GIPS firm."

Consistency doesn't necessarily make things right

When I was in the Field Artillery (many years ago), a gun battery would, once it had established a new position, orient their weapons to adjust for any misalignment which may have taken place since the last time they had been calibrated. This way, if a howitzer was consistently off, by adjusting for it the result would be as desired. A similar thing occurs with some golfers. For example, those who consistently slice the ball may aim to the left (for right handed golfers) of the fairway, knowing that their slice will cause the ball to go to the right, and hopefully in the fairway. Translation: in some cases, consistent errors can be dealt with to achieve the desired result; but this isn't always the case.

We generally want to see firms employ consistency when it comes to performance measurement. GIPS(R) (Global Investment Performance Standards) is one source for this desire. For example, firms need to be consistent in how they employ their rules for discretion, for their timing of adding and removing accounts, and for composite compositions. But if what you're doing is wrong, doing it consistently doesn't make it right.

Case in point. We have a client who employed a non-standard method to calculate performance; by "non-standard" I mean that it was invalid. The results would be often erroneous. Although they recognized the flaw in their method, they felt that by employing it consistently, somehow the error would be eliminated. Unfortunately this line of reasoning is invalid: consistently employing an invalid formula will only yield an invalid result.on a consistent basis (which might, in fact, mushroom!).

I do find it a bit odd that with all the documentation on how to derive returns, that individuals and firms will still come up with their own return method. And if they do derive their own approach, they should at least have it validated by an independent third party. Twenty-five years ago there wasn't a great deal that was readily available on how to calculate performance; today there are dozens of books and probably hundreds of articles, and loads of websites that can be referenced.

This wasn't the first time we've encountered firms who use their own formula. Often the methods that firms come up with seem quite intuitive; unfortunately, they're invariably wrong, however. What may seem to make intuitive sense often fails mathematically.

Personal returns ... how do they fit in?

I am conducting a GIPS(R) (Global Investment Performance Standards) verification for a client who asks their clients to provide a "personal benchmark." That is, the return the client hopes to get on an annual basis. Assuming there will be cash flows during the year, how does this factor into the asset management firm's reporting and measurement?

First, we must realize that the "personal benchmark" is a money-weighted benchmark; it can't be time-weighted. And therefore to provide for comparison purposes a portfolio return that's time-weighted is clearly a matter of "apples and oranges." What else could it be? And while this information has some value to the manager, they're managing to achieve a superior time-weighted return, working around the flows that come in and out of the portfolio. If the client makes some bad timing decisions which result in a "personal return" below their benchmark, who do we blame?

"Audited" returns

We are often asked to review a firm's performance measurement system,

• to ensure that the formulas are correct
• to possibly opine in writing on its value
• to identify opportunities for improvement.

We do this for software vendors, custodians, asset managers, and other service providers. And of late we've noted how some firms will refer to their returns as "audited." I have commented about the concept of "audited returns" from a GIPS(R) (Global Investment Performance Standards) perspective a couple times recently, but here I'm speaking simply of the returns which might be given to a client. Often service providers will provide their clients with daily or monthly returns, and will distinguish between "audited" and "unaudited."

It has occurred to me, as a result of the comments shared by CPA colleagues of mine and documented in the recent posts, that these terms are clearly inappropriate. Unless the provider brings in a CPA firm or some other outside body to "audit" their returns, they're technically always "unaudited." And, in fact, given the specific definition of what the term refers to, the numbers would never actually be "audited."

The appropriate word, I believe, is "reconciled." Is this not the case? A provider can, at best, reconcile the data that is used to produce their returns vis-a-vis the "official books and records." Perhaps "reconciled" doesn't carry the same perceived value as "audited," but it's more accurate, yes?

If a firm uses the term "audited," they should be prepared to explain what the term means. For example, "'Audited returns' means that we have reconciled the underlying data with the account's custodian, the source of the 'official books and records,' and have ensured that all securities are properly priced and ..." Do  you agree?