|
Home > Library > Stable Times > Volume 3, Issue 1
The quarterly publication of the Stable Value Investment Association
First Quarter 1999 • Volume 3 Issue 1
First Quarter 1999
Implementation of a Framework for Performance Calculations of GICs
By K. Daniel Libby, CFA Investment Manager IBM Retirement Fund
With the Stable Value Task Force's draft recommendation on Performance Measurement
for Stable Value Assets under consideration by the membership, attention
will shift toward implementation of these standards. The intent of this
article is to point out some of the issues that are outside the scope of
the Task Force's document but I believe will still need to be addressed
by the "participants" in this market, specifically Stable Value
Managers and Plan Sponsors. This article will also look at some of the alternative
calculations that are possible by various types of "participants"
to achieve the same performance measurements. In addition, it will comment
on the inputs required to provide these measurements in a consistent manner.
The Task Force concludes that the fundamental framework for this performance
standard should coincide with AIMR - PPS. That is to say that it is an "economic"
valuation approach and should seek to remove the effect of cash flow on
performance. Of tantamount importance in considering the implementation
of such a standard is that it be "Fair", "Robust" and
"Consistent". This article will discuss each in some detail as
it relates to implementation of this framework for performance calculations
for GICs.
Fair
Who "owns" the performance of the benefit responsive
insurance: Stable Value Managers or Policy Makers (Plan Sponsors) of the
Fund?
The "fly in the ointment" in applying such a framework to stable
value assets has long been the embedded benefit responsive insurance or
"puts". Why should this be so when putable corporate bonds are
handled in a straightforward fashion under AIMR - PPS? The reason, ostensibly,
has been that valuing these benefit responsive options is not a straightforward
matter. I believe, rather, that the difficulty lies in defining who "owns"
the performance of the benefit responsive insurance. Although I've made
this point in some of my previous articles for this publication, it is worthwhile
restating it somewhat differently for the purposes of this discussion on
implementation.
The analogy to buying putable corporate paper and GICs is a good one to
bring out this point. Putable corporate bonds are purchased as an economic
decision that the investment manager has authority over to exercise or not.
His decision to do so will be based on the relationship between interest
rates and the coupon rate on the putable corporate bond. Therefore the economic
performance of the value of that embedded put is very relevant to evaluating
the manager's performance.
A GIC is analogous to a putable corporate bond with the exception that the
manager does not have the authority to exercise the put embedded in the
GIC. It is the participants that will determine if and when to exercise
the puts based on a host of external factors such as demographics, plan
structure, competing investment returns and plan communications.
For this reason, I have long advocated that the performance standard should
not only be corrected for cash flow activity but also "grossed-up"
for the benefit responsive costs as well, at least for the purposes of external
publication and comparisons. This will have the added benefit of allowing
performance composites to be aggregated across various plans and allowing
performance comparisons among managers both within the stable value community
and outside it.
What are the practical implications of such a position in this framework?
As we know GICs are bond-like instruments that provide cash flows after
charges for benefit responsive insurance. Likewise, GIC rates are net of
these risk charges as well. Therefore, unfortunately, performance using
a discounted cash flow method for a GIC will reflect returns net of these
risk costs. Yet in today's market it is easy to obtain an indication of
the "global " wrap cost attributable to a plan and use this to
gross-up a performance return composite for traditional GICs that can then
be used for external publication and comparisons. As for synthetic GICs,
such as actively managed bond portfolios, performance net and gross of benefit
responsive costs is commonplace. Therefore, combining these return composites,
traditional and synthetic GICs, is straightforward and can be displayed
on a gross or net of benefit insurance charges basis depending on the intended
audience.
The larger implication of this conclusion is that it separates the strategic
(i.e. policy) decision involved in setting up the benefit responsive insurance
structure of a Stable Value Fund from the investment actions taken in managing
its assets. Whether the Plan Sponsor is involved in setting this policy
or delegates it to the Stable Value Manager does not change the appropriateness
of separating these two functions for the purposes of performance measurement.
Yet this implication should not be misconstrued as minimizing the role of
structuring, executing and benchmarking this policy. It is merely outside
the scope of this article to discuss it appropriately. But I would add that
the benefit responsive insurance structure lends itself better to consideration
as a policy decision not only because the investment manager has little
control over the exercise of those "puts", but also because the
insurance structure of the Fund overall typically cannot be rebalanced except
over very long horizons.
Robust
How should calculations be carried out for (daily priced)
Mutual Funds vs. (monthly priced) Stable Value Managers in an environment
where there are cash flows occurring into and out of the Fund?
Assuming that the portfolio manager does not "own" the performance
of the benefit responsive insurance, the calculation of performance for
GICs focuses exclusively on the discounted cash flow calculation. No additional
calculus is necessary for option valuation of the benefit responsive insurance.
In addition, since all that is necessary to know is the cash thrown off
by a GIC and the appropriate discount rate to use (see "Consistent"
below), no reference to the book value or the book-to-market ratio is required.
The AIMR Performance Presentation Standards Handbook, 1997 ed. (Chapter
2, pp. 43-51) provides all of the machinery necessary for these performance
calculations.
The AIMR-PPS discusses the Daily Valuation, Modified Dietz and Modified
BAI methodologies. While Daily Valuation is preferred, it may be impractical
for most participants other than mutual fund families that are organized
to provide daily pricing on its assets. For most stable value managers,
monthly valuations are standard practice and in those cases Modified Dietz
would be more appropriate. In a situation where the cash flowing into or
out of a GIC is significant due to a partial maturity or a deposit of a
contract, the asset should be re-priced just prior and just after the cash
flow and the resulting returns time-linked. This will remove the effect
of the cash flow on performance.
As for a "non-participating" GIC, specifically where a deposit
or withdrawal occurs in mid-period, the effect of the above or below market
contribution to the payment made by the benefit responsive insurance provider
is removed because it is priced into the basis of the asset both just prior
and just after the payment is made. An issue could be raised that the relative
importance of the GIC in question has shifted in overall market weight as
a percentage of the portfolio and therefore the effect of the cash flow
would impact the portfolio until the end of the period. However, typically,
this cash flow will be an insignificant proportion of the overall portfolio.
Therefore its impact in changing the market value weighting of the GIC in
question will also be insignificant. If it were not, a technique described
in the Handbook on page 49 as the Asset-Weighted and Cash Flow Weighted
Return could be used to mitigate its impact. It is interesting to note that
under Daily Valuation this effect is removed as well.
Consistent
How do you choose the correct spreads, adjusting for
differences in quote conventions, credit risk and optionality?
This is especially important for the results to be consistent to other markets
and across market cycles. While there is much that can be written to explain
the "correctness" of various methodologies to determine an "accurate"
valuation, that is not our focus here. For a performance framework the most
important consideration is that the practice be carried out in a plausible
and consistent fashion over time. In other words, the way GIC spreads are
chosen should be market-based and yet derived in a similar fashion each
reporting period throughout the life of the GIC asset. This will ensure
that the performance effect of the asset is properly represented in its
composite.
Conclusion
Each of the requirements that I have discussed as they
relate to implementation of a performance standard for GICs needs to be
addressed by the participants of this marketplace if they are to be successful
in their efforts. Other implementations are possible, however each will
have certain ramifications associated with them. My attempt has been on
providing a performance measure that is most analogous to existing performance
measures commonly in use by investors and their advisors and therefore allow
them to make the broadest use of their measure. It just happens that another
happy result of this framework is that it is quite straightforward to implement.
For both reasons but especially the former, I believe this is in the best
interests of this assets class and the participants in the long run.
Read Next: Interest Stable Value Yield Spreads
|
|
