Don't be afraid of "fair value"

Some GIPS(R) (Global Investment Performance Measurement) compliant firms are getting nervous about the upcoming requirement to use "fair value" pricing. For many, the concept is one that they historically saw no need to be concerned with, since they rely on market prices. But now they wonder what changes they will be required to make to maintain compliance.

The answer is quite simple: none!

Fair value is a general approach to pricing to ensure that the price represents the best price that one would expect to use to value an asset. GIPS provide a hierarchy to help firms navigate through the process to get this best price. And the very first level is market pricing! So, if you rely on market prices today and are able to get prices for all of your investments, nothing more needs to be done!

Aggregate method comments draw attention

I think I've gotten a record number of comments regarding this month's newsletter, some of which we'll share in the August edition. One reader encouraged me to transform it into an article, which I hinted that I might do and now have support, so I will pursue this.

I don't know how many vendors calculate their GIPS(R) (Global Investment Performance Standards) composite returns using the aggregate method, but believe that it's quite common. No vendor has reached out, yet, though I would suspect that some might. I wasn't attempting to cause any problems, but the evidence from my analysis does call into question this method's appropriateness.

Please let me know your thoughts, too!

Be careful when negative market values show up!

I'm reviewing a client's performance measurement system and came across an interesting example of how negative market values can wreak havoc with your reporting. What I'll be showing aren't exactly the numbers, but the problem is the same.

We begin the month with $20 and end with $10; we have no cash flows. And so what's our return? Hopefully you'll respond -50 percent. And what will their system show? Minus 180.46 percent! But how can this be?

The client uses a daily return approach and use Modified Dietz as their formula. They take the absolute value of the denominator in all cases. I want to pause here and say that while I've seen others employ this approach, it isn't clear that it's always a good idea, but with no evidence to refute its use, we'll conclude for now that it's okay.

The account in question is an interesting one in that it has cash and one security; the cash balance is negative for the entire time, while the security's price fluctuates greatly, causing it to go low enough on one day that the overall portfolio becomes negative on that day, meaning it swings from positive to negative and then back to positive.

On the day it begins negative and ends positive, it starts at -$3.00 and rises to +9.85. On that day, if we use the more standard version of Modified Dietz (or, what might be called the "true daily" formula: EMV/BMV), we get a return of -428.33 percent; if we use the absolute value in the denominator, we get +428.33. Now which makes more sense? If we go from a minus to a positive, wouldn't that seem to be a positive return? In this case only, I can see using the absolute value in the denominator as it provides (to me and many of my colleagues) a more reasonable and intuitive result).

Okay, so let's quickly look at our daily values and returns, as well as the return as derived by linking these daily numbers:

On June 23 we see the portfolio swing from a negative to a positive market value. If we link our daily returns we will get the -180.46% return shown at the top left; but if we calculate the return directly from the month's ending and beginning market values, we get a return of -50.00 percent. Also, if we don't take the absolute value of the denominator on 23 June (which results in a -428.33% return that day), then we WILL get a -50% return.

It looks like we have two choices: have a less intuitive daily return but a more accurate monthly, or a more intuitive daily but inaccurate monthly. But there IS a solution that resolves our difficulty.

The problem lies in how we link returns that come from short positions; we can arguably characterize our June 23rd starting position as being short, so we can employ it here. Traditional linking (i.e., (1+r1)*((1+r2)*...*(1+rn)-1) doesn't work with shorts. We have to subtract the return from 1 on the date with the short position. In our case, it will change the value from 5.2833 to -3.2833. And, when we link it to the other daily values we achieve the -50% return we would expect.

I often say that the move to daily returns actually creates more challenges to accurate returns; this is just one example. I will comment on this in greater detail in next month's newsletter.

The way we always do it isn't necessarily the way to do it

What a blog title, huh?  Anyway, you'll soon get my point.

In last night's game against the Cleveland Indians, with men on first and second and one out, the Minnesota Twins did what they're supposed to do: move the runners along. And so, they had their catcher, Joe Mauer, lay down a sacrifice bunt, which he did successfully, moving the runners to second and third, so that both were now in scoring position. And what did the next batter do? He made the third out.

Perhaps the Twins' manager should have read Michael Lewis' Moneyball: The Art of Winning an Unfair Game. Lewis pointed out how the Oakland As went against "baseball wisdom" and managed to do quite well. And one of the things they avoided: taking the bat out of a good hitter, like Mauer. It turns out that more often than not, the sacrifice bunt strategy fails. But why do we do it? Because that's what you're supposed to do!. But who says?

The same problem arises in many other aspects of life, where someone (who knows who) decided that there's a certain way to behave. And in performance measurement we see it all the time, perhaps most noticeably in the overuse of time-weighting. Too often we go along with the crowd because it makes us safe. If the Twins manager last  night had Mauer hit, and he hit into a double play, many would have said "well, he screwed up: didn't he know he was supposed to move the runners along?" But since Mauer is a .300 hitter, allowing him to hit would have been the right move, just as money-weighting is more often than not the right move.

Linking fixed income attribution

As you're probably aware, arithmetic attribution results in subperiod effects which are linking challenged; that is, if we attempt to link these subperiod (e.g., monthly) results to obtain longer (e.g., annual) period results, using, for example, simple geometric linking (as we use with returns), the results won't reconcile to the longer period excess return: a residual typically results [geez! what a sentence!]. As a result, several individuals (e.g., Jose Menchero, David Cariño, and Andrew Frongello) developed models which eliminate the residuals. These models are typically demonstrated with one of the Brinson (i.e., Brinson-Hood-Beebower and Brinson-Fachler) models.

But what if you're doing fixed income attribution and using, for example, the Campisi model: how do we link these subperiod results?

Answer: the same as we would with equity attribution; that is, use one of the linking models. Special models aren't required to make this work.

"Men have had recourse to many calculations"

The sources of inspiration for blog pieces seems almost infinite. Today, while reading the bible, I found the above quote (it's from the book of Ecclesiastes, Chapter 7, verse 29). Men (and women), in general, obviously have many calculations to draw upon; and even we in performance measurement have many: the challenge is to decide which ones apply!

I guess that's the beauty of what we do, as it provides us with flexibility and options.

The problems arise when individuals use erroneous calculations or use them in a manner that is inappropriate. And some firms have been known to derive their own calculations, even though there are plenty good ones already available: and unfortunately, sometimes the ones they derive are flawed!

We will always have debates about the various approaches (e.g., geometric vs. arithmetic, transaction-based vs. holdings-based), but that's fine. "Variety's the spice of life," yes?

Don't reconcile daily? Don't worry ... what's an error here and there?

The Global Investment Performance Standards (GIPS(R)) now require compliant firms to revalue their portfolios for "large cash flows." But what does "revalue" mean?

Does "revalue" simply mean "reprice"? If it does, then that means that on a large flow day you might actually not have the correct holdings in your portfolio, but the repricing will be deemed sufficient, even though the resulting market value is in error.

If it does mean "reprice," then why not simply call it "reprice"? That would eliminate confusion.

I would argue that revalue should mean that you literally revalue the portfolio, meaning that you reconcile the positions to ensure that they're right, and then reprice.

Firms that don't reconcile on a regular basis run the risk of having errors. Most managers no doubt reconcile only monthly, which is fine, provided they make adjustments to prior periods, as necessary; especially if they're either doing daily performance or revaluing for large flows.

Make sense?

Does the world revolve around GIPS(R)?

How many performance standards can you think of? Well, we have (of course) GIPS(R) (Global Investment Performance Standards), the AIMR-PPS(R) (no longer valid), UKIPS (no longer valid), several other country standards (which are no longer valid), the BAI (Bank Administration Institute) standards, and the ICAA (Investment Council Association of America; now the Investment Adviser Association)) standards.

But when people generally think about performance standards, what usually comes to mind? Chances are your answer will be GIPS, right? And so, many folks who want to know what the "rules" are will turn to GIPS, even though they don't necessarily apply.

We see this in so many sectors of our industry, from brokerage to pension funds, from custodians to consultants. The question "are you compliant?" when directed to custodians and software vendors has long been identified as an inappropriate one to pose, but pose it many people do.

We're doing a project for a service provider whose clients are asking if their fee calculation is "GIPS compliant." Well, it is and isn't, but that's not important: what's important is the question, "why does it matter?" The output from this provider isn't used by asset managers. Granted, GIPS may often be seen as "best practice," but their rules don't always apply.

"When the only tool you have is a hammer, all problems look like nails"

Life in performance can be confusing, yes?

Back (or) to the future?

We are hearing from firms on a regular basis who want to become compliant with GIPS(R) (Global Investment Performance Standards). More and more hedge funds are showing interest, as well as long only managers and others. While we attribute part of this sudden interest to Bernie Madoff, no doubt the market downturn has provided motivation, as well as the desire for many of these managers to enter the institutional space, where compliance is virtually mandatory.

And so, what set of standards should the firm comply with, the 2005 or 2010 edition?

Clearly, 2010; why bother going through all the effort to comply with a version which is about to become obsolete. Get a copy of the new version and comply with that. You'll be able to still use much of what's in the current version of the handbook, too. Don't look back, move forward!

Accuracy vs. Consistency

I'm conducting a review of a service provider's performance measurement system. The issue of as-of adjustments came up and they explained that their clients prefer not to see their returns revised because they'd then have to explain why they're different. Is it that they don't want to see change? I guess to some extent, this is the case. The client summarized this by saying that they're clients:

"want consistency over accuracy."

Their clients

"don't want to explain the differences."

I do consider this unfortunate, because we can liken it to the ostrich who buries their head in the sand; they don't want to know what's really going on. 

I wonder how common this is? The GIPS(R) (Global Investment Performance Standards) now require compliant firms to have an error correction policy. Firms can, of course, choose a very high level as being "material," which could obviate any need for making corrections. But what's wrong with correcting something that's been updated? Surely firms know that there can be errors! Oh, well. Perhaps there's something to be said for consistency ... just not sure what it is.

Balanced composites ... are they needed?

What is a "balanced manager"? Arguably, they're a manager who (a) claims some expertise in managing two or more asset classes (e.g., stocks and bonds) and (b) has an ability to adjust allocations across these asset classes. Well, what if the manager doesn't control the allocation? What if it's the client who dictates the allocation? In these cases one might argue that the manager isn't actually a "balanced" manager, but rather a stock and bond manager.

I base this statement partly on a perhaps long forgotten document titled "Answers to Common Questions About AIMR's Performance Presentation Standards," dated September 1992 and published by the Association for Investment Management and Research (what is called the CFA Institute today). Here we find the following:

• Question: "If the client dictates the asset allocation in a balanced portfolio - e.g., 40% equities, 60% fixed income - how should the portfolio returns be reported?"
• Answer: "If the client dictates the mix, then this portfolio is not a balanced portfolio, because the manager does not have any discretion over asset allocation, which is a main component of return for this strategy. The Standards recommend that these segment returns be included in equity-only and fixed-income-only composites, with cash accurately allocated."

I should mention that with much of what we do, not everyone agrees with this statement. Many situations involve a client who provides the manager with a range, where the manager has discretion to allocate within certain boundaries. But even here, one might argue that the client has taken responsibility for allocation.

The bottom line is that firms can approach these situations in a manner they feel most appropriate. Some firms will treat these accounts as "balanced," although they aren't the one who is "calling the shots" on the allocation; others will declare them separate asset classes that need to be composited within their respective asset class specific composite.

We began working with a new verification client who has a massive amount of high net worth clients who in all cases dictate the allocation: I suggested to them that they don't have to create a multitude of composites that cover all such ranges. Ideally, they should "carve out" the equity and fixed income pieces and put them into their appropriate composites. However, this ability has been made more challenging with the January 1, 2010 requirement that cash be maintained separately. Alternatively, the firm could declare the accounts as being "non-discretionary" (for GIPS(R) (Global Investment Performance Standards) purposes) since, as the AIMR document clearly states, the allocation drives a "main component of return."

Why time-weighting if we don't weight time?

There seems to be some confusion as to what "time-weighting" actually means. The term was coined in the 1968 Bank Administration Institute (BAI) standards. The BAI proposed three ways to calculate returns:

• the "exact method," whereby we revalue the portfolio for any cash flow
• the "linked IRR," where we geometrically link subperiod (e.g., monthly) returns which were derived using the internal rate of return (IRR); this is similar to the Modified Dietz formula
• the time-weighted method, where instead of geometrically linking, we link returns based on the length of time between flows.

The third method can produce returns which are in error, thus it's been abandoned. However, the term "time-weighting" remains in our lexicon. But do we "weight" time? Nope! In fact, time has no bearing whatsoever on our returns. If we link a one day return, with a one week, one month, one quarter, one year, and one decade return, we will obtain a cumulative return across the full period, but in no way do we give extra "weight" to any of these periods.

In no way do we weight time in time-weighting. Granted, we weight cash flows based on their time in the Modified Dietz and Linked IRR, but this wasn't the source of the expression. Time-weighting simply means that we are eliminating or reducing the impact of cash flows. That's it! No time weighting.

Much ado about nothing?

A colleague, when seeing my commentary on the GIPS(R) (Global Investment Performance Standards) use of the aggregate method to derive composite returns, stated "I don't think this is an issue we should worry about ." Suffice it to say, I disagree.

Why did the standards introduce a requirement this past January to require firms to at least revalue for large cash flows? To provide more accurate results. That's it. And so, many firms have had to "beef up" their calculation process in order to comply. Great!

But then, many of these firms employ the aggregate method to derive their composite results, which are going to take these highly accurate portfolio returns and turn them into less-than accurate composite returns! Does this make any sense? And, for the standards to require managers who use the aggregate method to revalue when these flows are the result of new accounts added within the period will generate even less accurate results.

My colleague also suggested "Basically accounts should not enter mid-period - and if we don't say that anywhere perhaps we should."  Funny: suggesting that we disallow the entry of new accounts mid-month but to not support stopping the use of a composite return method which is clearly flawed!

Perhaps I am alone in this subject as we've permitted the aggregate method for 17+ years. But, if we stopped allowing it, would there be a problem? We stopped permitting the Original Dietz method, which was once permitted. If we're striving for accuracy, why would we permit an admittedly flawed method to continue to be used? Am I tilting at windmills?

More on the aggregate method

When I posted yesterday's blog entry, I was only fixated on a simple issue: revaluing a composite for large cash flows from new accounts when using the aggregate method. The GIPS(R) (Global Investment Performance Standards) permit the use of the aggregate method, along with two asset-weighted methods, to derive composite returns. And a recently published Q&A on the GIPS website states that the 1 January 2010 requirement to revalue portfolios for large flows also applies to composites when deriving returns using the aggregate method. And while I didn't entirely object to this requirement, I do in the case of new accounts being added during the period, as noted in yesterday's post.

However, I have begun to think about the aggregate method a bit more, as I said I would. I have concluded that there are three issues worth considering:

1. When using the aggregate method, should the firm revalue when a new account is added during the period? I think that I demonstrated that doing so can result in erroneous and/or misleading results.
2. When using the aggregate method, should (as required by the GIPS Q&A) the composite be revalued for large flows (other than for new accounts)? Here, I believe there is evidence to suggest that by revaluing the result will be more accurate, and so I'd say "yes."
3. Should the aggregate method even be permitted? Or, if permitted, should caution be given in its use? Clearly, this is a much more sweeping question and I have concluded that this method is potentially flawed, both in theory and in reality and should no longer be permitted.

A number of years ago I was doing a talk in Europe on the various attribution linking methods: I presented the Menchero and Cariño models, and then a somewhat naïve approach, which was much more intuitive. Jose Menchero was in attendance and after my talk asked to give him 24 hours to disprove the naïve method. And sure enough, the following day Jose gave me an example which, when using this method, yielded a counter intuitive result. In science, all we need is one example to disprove a theorem.

And we have such a case with the aggregate method:

In this example we have three accounts, each of which experienced a 4.00% return for the period. Halfway through the month, the third account had a 30% cash flow. At mid period, you can see that one account was up 1%, one was down 1%, and the third was even. If we revalue for the flow, the aggregate method yields a return of 4.10%; if we don't, the result is 4.29 percent. Do either make sense?

The composite return is supposed to reflect an average of actual results: but the actual results were all 4.00%; it appears that our aggregate method generated composite return overstates what actually occurred. Both asset-weighted methods produce correct results. No one is managing the composite, so what does it matter how the composite did?

A colleague, whose opinion I value greatly, offered the following:

"if you treat the composite as if it is one big portfolio 

then its only logical that the cash flow rule applied to that composite as if it were a single composite"

I totally agree...it is only logical that the cash flow rule should apply.  But in this case, logic fails.

In our July newsletter I will provide additional examples which call into question the aggregate method's efficacy as a composite return method.

Revalue portfolios, not composites, for large cash flows (at least when the flow is a new account!)

Effective 1 January 2010, GIPS(R) (Global Investment Performance Standards) compliant firms are obligated to revalue their portfolios for large cash flows. A Q&A made it clear that this applies to firms that use the "aggregate method" to derive their composite returns, even though the underlying portfolios aren't used in this method to come up with the return. However, I would argue that it would be inappropriate to revalue composites themselves for large flows! I first addressed this in our November 2006 newsletter, following a project I conducted for a unit of Bear Stearns (I'm comfortable mentioning the client now, given their demise). The project involved a case where they grouped accounts together to provide a client with an overall representation of their performance. And while I later argued for the use of money-weighting in these cases (not for GIPS purposes, but for client reporting), I nevertheless tackled the time-weighting angle. It seems that they replaced a prior method with one that revalued the collection (aka "composite") when large flows occurred. At first, I thought this was probably acceptable; however, the results proved nonsensical, which caused me to reflect more about this.

I raise this issue again because (a) it has been some time since I first tackled this, (b) GIPS now requires revaluation for large flows, and (c) a (new) client presented a problem they were having with returns which smacked of this exact issue, and so I will repeat myself a bit, so as to reaffirm the harm that can result from revaluing composites for large flows.

Here's the scenario as laid out in the November 2006 newsletter:

Pretty simple case, right? We have a composite that begins with a single account, which is joined half way through the month by a second account. Should we revalue the composite because of the large flow?

Well, if we do the result will be 1.33 percent. Does this make any sense? After all, the two portfolios have returns that both exceed this value, so how can this be right? When I was presented with an example like this from Bear, I was at first perplexed, because I couldn't see why the result would be so obviously wrong!

If, however, we use an asset-weighted approach that avoids revaluation, the result is the more plausible 3.33 percent! I will leave it to the reader to either validate these results on your own or to visit the earlier newsletter which provides the details.

And so, how can we explain what's happening? Let's revisit the reason we revalue: to eliminate the impact of the cash flow on the portfolio, so as to separate the timing before and after the flow event. However, is the manager managing the composite? Hardly! They're managing two separate portfolios. To revalue the first portfolio because a new one is introduced causes us to capture a point where the account had dropped in value: but what does this have to do with the second portfolio being added? Nothing!

Firms that go to the trouble of revaluing composites for large flows are opening themselves to reporting erroneous results when they revalue for new accounts being added during the month. Granted, most firms bring on new accounts at the start of a new month, so this wouldn't be an issue, but for those that will bring an account in mid-period, problems can arise and arguably any result is an error.

Should the firm revalue for flows other than for new accounts? I'm not sure about this as it would require more thought on my part. But for now, let's say I'm uncomfortable with the notion of revaluing an entire composite because one of the accounts has a very large flow: I don't believe the result will necessarily be accurate.

p.s., I want to point out that the result I obtained without revaluing was by using the asset-weighted method, not the aggregate method. I am exploring the impact of revaluing vs. not on the aggregate method and will post something shortly.

p.p.s., I also want to emphasize that my current position about revaluing is regarding cases where new accounts are added to a composite during the period; I am researching the broader case of revaluing.

A balanced approach to risk

I typically tell the students who attend our Fundamentals of Performance Measurement course that risk is the most challenging aspect of our industry. We can't agree on what it means or how to measure it. And while many today see risk as either being the potential for loss or inability to meet an objective, most of our risk measures actually measure volatility.

Because risk is so difficult to get ones arms around, it needs to be approached from multiple directions, so as to get the best picture possible of what we're facing. In addition to typical risk measures such as standard deviation, tracking error, and beta, and risk-adjusted measures such as Sharpe ratio, Information ratio, and M-squared, we should consider other aspects of risk, too.

In spite of its challenges, Value at Risk (VaR) provides us useful information. Liquidity risk is another measure one should be assessing. At last month's PMAR Europe, Jose Menchero discussed "extreme risk," which can be seen as a complement to VaR.

Failing to properly assess risk is risky: looking at it from multiple angles is one way to gain some insights into what's going on.

Happy July 4th!

Palm Beach, Florida-based money manager Carl Domino, Inc. typically sends notes preceding various holidays, and this week was no exception.

I was one of the many fortunate ones who received this greeting:

July 4th is an important day in America’s history.
A day where Americans celebrate their freedom, 

appreciate their country, display our nation’s flag 

and pay respect to those that died to protect us.

I can't say it any better.

On behalf of The Spaulding Group team, we wish you a happy and safe July 4th!

Wrong index or active management?

I taught our Introduction to Performance Measurement course yesterday and one student asked how one might know whether, if the tracking error is high, it is because of (a) active management or (b) the wrong index being used?

I don't believe there's a simple way of knowing this. If the tracking error is exceptionally high (e.g., 15%), then the manger either obliterated the index or is doing something entirely different. So perhaps some levels make it clear that it's probably the wrong index. Style analysis could be a tool that can be used to analyze the portfolio and see if it aligns with the index, at least to some extent. A review of the holdings, sectors, market caps may also be in order. If it's a blended index, does the blend include all sectors in which the manager is investing?

We occasionally see managers use broad indexes, such as the S&P500, when their strategy involves a single style (e.g., growth) or market cap (e.g., mid cap). In this case, it's pretty clear that we have the wrong index.

If you have ideas about this, please chime in. Thanks!

When is an account, not an account?

I just got a question from a client. They often bundle accounts into a "household" for reporting purposes. And, it's the household which actually reflects their strategy; that is, the individual accounts may have pieces of the strategy, but the actual strategy is acted upon at the household level. And so, how do we handle this for GIPS(R) (Global Investment Performance Standards) purposes?

In this case it would be appropriate to treat the household as the account, even though it may not legally be an account; it's a "virtual" account, yes? THIS is what you're managing, and you're using the pieces because individually you may not be able to execute your strategy. And therefore, the pieces wouldn't be in a composite by themselves, but the entire household would be. Make sense?

And, the firm's policies and procedures should reflect this, too!