We are hosting a "mini" survey, and would like your thoughts, IF your firm claims compliance with the Global Investment Performance Standards.
There are really only two questions. It should take you only a minute or so to complete it!
The first deals with sunset rules.
The second is about the use of Q&As to establish rules.
As an incentive to participate, we will have a drawing for a $25 American Express gift card!
The responses will be detailed here, as well as in our newsletters. Please visit our survey site, and submit your answers by April 21.
Thanks!
A client sent us a question that I am a bit surprised hasn't been asked before: which country's risk-free rate should be used when investing across countries?
For example, if a US domiciled investor has a Japan-based asset, when we calculate the Sharpe ratio, should we use a US risk-free rate or one from Japan? Likewise, if we have a client in Japan for whom we've purchased assets in the UK, which country's risk-free rate should we employ?
In risk-adjusted return measures such as the Sharpe ratio, we use the risk-free return to derive the risk premium (portfolio return minus risk-free return). It's the premium the investor is entitled to, for taking on more risk.
If the portfolio manager picks a Japanese asset, presumably it's because they want to gain exposure to that market. And so, to me, the alternative risk-free asset would be one from Japan.
A challenge arises when we invest in a country (e.g., emerging market) where there is no risk-free asset. In these cases, I would think it reasonable to default to the risk-free asset of the investor's home country. These are views that I've come up with without the benefit of discussion with others, and so I am open to being corrected, enlightened, persuaded to adopt alternative ones.
This is an interesting topic, I believe, that is worthy of much discussion. This is the first time I've opined on it, and am curious what you think, so please offer your comments!
Recall that the GIPS(R) (Global Investment Performance Standards) Executive Committee (EC) issued a "Q&A" that introduces a change to the Standards, allowing firms to remove disclosures of composite name changes after five years.
During the recent EC open meeting call, I asked if it's possible to see additional GIPS requirements receive similar "sunset provisions." The response was, as one might have expected: "yes, of course it's possible." I would hope we could do better, though.
I remain of the opinion that Q&As are not the place where changes should be introduced. There's a clear precedence for them to be put forward for public comment. We know that President Barack Obama has been criticized and accused of bypassing congress, by making changes to laws (only time will tell if this is formally addressed); the EC's practice of using Q&As as a way to avoid putting new rules out for public scrutiny is one that I believe should cease, unless those very same Q&As are also put out for comment (which actually wouldn't be a good idea).
I received an email this week from a verification client asking if they can drop some of the disclosures which to them, and probably just about everyone on the planet, are no longer relevant. However, in the absence of an official change, they cannot. In an earlier post I suggested that since this change was introduced through a Q&A, perhaps the ability to make "judgment calls" applies everywhere, but prudence suggests that one should be circumspect about such actions.
Given that the EC has decided to cease the practice of quinquennially reviewing and introducing new versions of the Standards, one would think that time is available to take a serious look at all the required disclosures, and determine which, like composite name changes, can be permitted to disappear. A "Guidance Statement on Sunset Provisions" would be welcome by all, I believe.
We will shortly conduct a "mini survey" on these topics, to gain some insights into what others think.
Of late, my colleagues and I have been involved in numerous conversations on the subject of benchmarks; in fact, I'm about to conduct a focused study for a client on their benchmark construction policies and procedures.
I conducted a GIPS(R) (Global Investment Performance Standards) verification for a firm who occasionally uses benchmarks solely for "reference" purposes; this is quite common with hedge funds, that are "absolute" strategies, where no appropriate benchmark exists. And so, an index like the S&P 500 might be used.
If we look at the definition of "benchmark" in the Standards' glossary we find:
At first blush, one would be inclined to believe that such a benchmark is sufficient, and fulfills the requirements of the Standards. Sadly, this definition is incomplete and therefore misleading, thus the confusion that can arise.
Within the corpus of the Standards we find:
THIS, to me, is a much better definition. Hopefully, in the 2020 or 2025 version, we'll see an expansion of it in the glossary.
And so, if the benchmark is solely for reference purposes, although it fulfills the definition, it fails to meet the requirements of this provision (i.e., I.5.A.1.e.), and the firm must include a disclosure, as required by:
and the "reference" benchmark should be flagged as "supplemental information."
Simple, right?
We received an inquiry recently regarding the appropriateness to rebalance a benchmark when the portfolio has a 10% flow. I discussed this with my colleagues, and we were all a bit surprised by the question. Now, perhaps there are loads of firms that do this, and we're just not aware; or, the questioner was confusing rebalancing benchmarks with revaluing portfolios to calculate time-weighted returns (a common practice to revalue a portfolio when large flows occur).
The benchmark represents the strategy. What should cause it to be rebalanced?
Well, let's consider the portfolio and benchmark at the start of investing. Let's make it really simple, and say that there are two asset classes, stocks and bonds, and that the strategy is to be 50/50 (i.e., 50% stocks, 50% bonds). Almost immediately, in response to market movements, both will drift from their initial allocation. And even though the strategy is 50/50, the portfolio may have a different one, based on tactical decisions of the manager; perhaps the manager decided to overweight stocks, so the portfolio begins at 60/40.
And so, we see that they both drift. Let's project that stocks are doing much better than bonds, so that after a month the benchmark is 55/45 and the portfolio's at 65/35. Do we rebalance?
To rebalance the portfolio, the manager must sell some stocks (that appreciated) and put the money into bonds (which didn't do as well relative to stocks during this period). This trading is costly, yes? Transaction costs (e.g., commissions) must be paid. A manager may elect to rebalance in order to reduce risk: while the portfolio benefited from the market, the farther away it goes from its initial target, the greater the portfolio's risk, should the market turn. And so, the manager may forgo further increases to the full amount currently sitting in stocks, in order to protect the portfolio from a jolt in the market. Some managers will allow the portfolio to continue to drift, however. Let's say that the manager does not, at this point, rebalance the portfolio; should they rebalance the benchmark?
Unlike rebalancing the portfolio, when we do this to the benchmark there is no cost; it's just making adjustments to the weights and calculating the appropriate return: no transactions are needed and no transaction costs are incurred. So, should the manager rebalance the benchmark?
There are at least two options:
#1 Since the portfolio is allowed to drift, to bring the benchmark back into alignment would create a situation where the portfolio is doing something the benchmark isn't allowed to do, and so perhaps it's gaining an advantage.
#2 HOWEVER, recall that benchmarks shouldn't include tactical decisions. In addition to an allocation that's different from the strategy, I'd say that to allow the allocation to drift (in response to the market) is also a tactical decision. Let's say that the equity portion continues to grow, and then BOOM, we get hit with a big market adjustment; the portfolio's return will drop. Well, if we allow the benchmark to drift, too, it will suffer from the same market downturn. But the benchmark is no longer the strategy, it's drifted away. And so, I would say that "best practice" would call for regular periodic rebalancing of the benchmark (most likely monthly), to ensure it remains representative of the strategy.
And so, in answer to the question, "what triggers a benchmark rebalance," I would say time.
Another component could be drift, right? The firm can establish a threshold that says if the benchmark's strategic allocation shifts by X% or more, it will be rebalanced. I'd think this would be fitting, too.
The benchmark is to align with the portfolio's strategy, and one would think that the rebalancing rules should be a component of the strategy: if it calls for quarterly rebalancing, then the benchmark should rebalance at the same point. If the manager decides to override the strategy and use a tactic to gain an advantage, then I would think the benchmark should continue with the strategy's rules.
This is an area where I haven't done much dabbling, and so don't know what common practice is, or even if there is a "common practice," so am open to your thoughts, suggestions, ideas, etc. Thanks!
When we conduct GIPS(R) verifications, it is not uncommon to see a footnote appear on a presentation where the number of accounts in the composite are five or fewer. The Global Investment Performance Standards do not require firms to disclose the number of accounts or dispersion if there are five or fewer accounts (at year or end or for the full year, respectfully), and most firms avoid doing it. Footnotes often include wording such as "statistical measures of internal dispersion are not considered meaningful and therefore not presented."
Not meaningful. What is meant by that? Statistically significant? Doubtful.
If there are six accounts present for the full year, is dispersion "meaningful"? Apparently, yes, though I believe statisticians would probably argue that they are not. For standard deviation to have value, we'd probably want 30 or more observations, which is a threshold that is well beyond what is permitted.
Simply indicating "≤ 5" should suffice for number of accounts, and "n/a" for dispersion, though a footnote indicating that there were less than six accounts present for the full year will work, too.
I delivered a talk earlier this week for the CFA Society of St. Louis on performance attribution. One of the attendees asked "what's the point?" I.e., why should an asset manager be doing performance attribution?
My initial response was "every party has a pooper, that's why we invited you," but then observed that this was actually a very good question. Why bother?
I explained that while it is a great tool to explain how a manager outperformed their benchmark, it can be even more beneficial to explain why they underperformed. No one wants to be told "I don't know," or to be given a vague excuse. If the manager can delineate the basis behind under-performance, it clearly and loudly communicates that the manager has their finger on the pulse; that they understand a great deal about their process and the market.
We have spoken with several individuals who have testified to the benefits such reporting has, and how it has helped them retain clients that they may have lost had they not had such information at their disposal.
And so, it does matter ... a lot!
Although not common, I still occasionally run into situations like the following: a period is selected to report performance that extends beyond (at the start, end, or both) the actual period a portfolio or asset was being managed or held, and yet a return is produced, with no indication of the true period for which the return is being measured.
For example, a portfolio's inception date is October 20, 2013, and you're reporting the fourth quarter 2013 returns for all portfolios. This portfolio, along with every other one in the composite, has a return, with no flagging or footnote indicating that its true period is shorter. To me, this is a big problem.
We occasionally see this when we conduct GIPS(R) verifications, where a composite began after the start of the year (or ended before the end, or had a break within the year). We require our clients to have a footnote or some other indicator as to what the true reporting period is.
It occurred to me that in cases like this, we're reporting a return that occurred during the period, but it's definitely not for the period.
Imagine a prospective employee who shows you their resume that reflects a continuous employment history, with no breaks. You later find that they had been terminated from one job and that it took them six months to get a new one. To avoid the unseemly break, they decided to extend the termination date from the one job, and move the start date from the second job up a bit. And so, they're reporting on jobs that were held during the alleged period, but clearly not for the period. We would not look favorably upon such reporting, would we? Folks have been terminated for less.
Well, shouldn't performance reports be held to the same level of scrutiny? My belief is that if a portfolio, security, etc. doesn't exist for the full period, you either (a) report "n/a," or (b) report a return with an indicator that it's for a shorter period, and identify what that period is. To do otherwise would, in my view, provide invalid and misleading information.
Any thoughts?
.png)
