 |
|
||||
| Home | News | Help Desk | Membership | Library | About |
|
| Library | |
 |
|
| Stable Times | |
| Papers | |
| Fee Disclosure Template | |
| Key Principles | |
|
Home > Library > Stable Times > Volume 3, Issue 3  The quarterly publication of the Stable Value Investment Association Third Quarter 1999 • Volume 3 Issue 3 Are GICs Efficient Assets to Include in Stable Value Funds?By K. Daniel Libby, The IBM Retirement Fund "Show me the risks..." Stable value funds, in many ways, are the simplest of investment vehicles. Participants seem to understand these funds very well. Analogous to passbook savings accounts, they provide the return of bond funds with the volatility of money markets. This seemingly incongruous set of facts gives these funds risk/return profiles that appear too good to be true.Stable value investment professionals know that this, in fact, is possible and that it results from the benefit responsive insurance within these portfolios. Furthermore, they will tell you that there is no "free lunch" in these vehicles; the price for this risk reduction must be tallied so as to include both the explicit as well as the implicit costs, i.e., participant transfer restrictions. However, the unfortunate truth is that while these funds may be simple for participants to understand there is little true understanding of how these funds work outside of the stable value community. More pointedly, there is little understanding of the risks undertaken in these funds. In fact, the appearance of the absence of risk may only exacerbate the problem. Indeed, there are many types of risk involved in managing a stable value fund. These include liquidity risk, credit risk, duration risk, sector allocation risk, active/passive risk as well as the risk of non-competitive returns. But these are much the same risks that exist in all fixed income portfolio management. So while the benefit responsive insurance is responsible for "simplifying" these funds for participants, it is also responsible for obfuscating the risks underlying these funds to professionals and participants alike. The adoption of a performance measurement standard for stable value funds will alleviate this issue. In traditional portfolio investment management, AIMR-compliant Performance Presentation Standards (PPS) are used to quantify these same risks. For example, a manager's return stream can be evaluated many different ways to detect a manager's expected alpha as well as any style bias. Examples of style bias would include holding excess liquidity, lower credit-quality assets, duration and sector bets, just as we mentioned above as being risks currently hidden in the returns of stable value funds. Yet this is only half of the problem. Even with the proper measurement of a GIC portfolio's performance, the fundamental questions still remain. "Are the GIC holdings being utilized efficiently?" "Do GICs represent a competitive asset to traditional fixed income securities?" In other words, how will we benchmark the use of GICs in a stable value portfolio? Drawing informed conclusions about the risks inherent in a portfolio based on a stream of performance numbers requires a baseline for comparison. Currently, there does not exist a GIC sector index as a component of a Lehman-stlye family of indices showing periodic total rates of return; such as exists for other fixed income asset classes. If such an index were available, these questions, and many like them, could be addressed directly. In lieu of that, we can try to answer these questions by creating a surrogate GIC index showing historical total rates of return. A Surrogate GIC Index First, a caveat is in order. Anytime there is an attempt to recreate history; one must understand the limits of the data they are working with. The construction used here is meant to quantify, generally, the performance and risk characteristics of GICs historically. Data was gathered from multiple data sources, namely Bankers Trust and T. Rowe Price; and although every attempt was made to ensure consistency, in the end there is no substitute for actual performance data from a single source. The author would like to thank John Hancock's Wayne Gates and Klaus Shigley, respectively, for assisting with the data used for this analysis and for providing their comments to the analysis. However, any errors or omissions in the use of this data or the conclusions drawn from it, unfortunately, are my own. While the GIC marketplace does not have readily available total return series', it does have a fairly lengthy history of GIC spreads at various maturity points. We begin by accumulating GIC rates for 2 year, 3 year and 5 year maturities going back to January 1983. Again, some care must be taken to maintain consistency due to various quote conventions for different data streams from different providers. Nonetheless, recognizing that if we assume the initial economic value of a GIC is par then, given the yield, coupon, maturity and change in yield, it is possible to determine its total rate of return. In this way we can use the GIC offering yield each month to compute the total rate of return of a constant holding in a given maturity of GIC assets. For example, the time series of 2-year GIC rates can easily be converted into a total rate of return time series for an investment strategy of continually selling last month's on-the-run 2-year GIC and purchasing this month's on-the-run 2-year GIC asset. Note that here we are assuming that the offering yield can also be used for the bid pricing at month-end. But, similarly, traditional fixed income indices are passive, buy-and-hold, portfolios priced consistently at one side of the market to minimize transaction costs as well. As we compute the total return time series for 2-year GICs we can also compute the duration of this simple portfolio. Likewise we can calculate the same information for the 3-year and 5-year time series' as well. Armed with these three data streams across the maturity spectrum we are ready to construct portfolios of GICs as linear combinations of these three investment strategies to target any desired duration between approximately 2 and 5 years. Duration-weighted GIC portfolio yields are available as well. The purpose of modeling a GIC portfolio by using three maturity points is to try and capture some of the yield curve exposure that an actual portfolio would have. The final step is to choose constant weightings of the three maturities of GICs to create a GIC portfolio to compare against a public fixed income index. As for the choice of public index, it would appear to make sense to evaluate two investment-grade indices against this surrogate GIC index: the Salomon Brothers 1 to 5 Year Government/Corporate index and the Salomon Brothers 1 to 5 Year Corporate index. These indices will straddle the quality rating of a portfolio of GICs. Furthermore, the absence of option-laden mortgage holdings in these public indices also would be appropriate for evaluating a traditional GIC portfolio. Several techniques were explored to calibrate the weightings of GIC maturities within the GIC portfolio, including using regression analysis to maximize R-squared. These did not change the conclusions discussed here. The technique used below to calibrate these weightings was to match the average index duration since January 1983. This resulted in a 60% weighting in the 2-year GIC strategy and a 20% weighting each in the 3-year and 5-year strategy. The resulting comparison of the yields is as follows:
To get a better idea of the efficiency of GICs, a closer look at the returns generated by these indices follows:
Annualized Excess Return = +28 bps.
Annualized Excess Return = -29 bps It is worth noting that the framework employed here omits any adjustment to the surrogate GIC index for realized default and loss experience in the early 1990s. While no definitive study has been done, this author is aware of some analysis by industry professionals. Again, while data is sketchy and interpretations may differ, a reasonable estimate for GIC annualized loss experience appears to be under 4 bps, perhaps considerably under. In any event, omission of this loss experience's impact on return does not change the validity of the analysis performed here. Conclusions The question of efficiency in asset pricing revolves around economic rates or return for equivalent levels of risk. Risk can be thought of as being comprised of several dimensions: liquidity, credit, sector, duration, active/passive and the risk of non-competitive returns. This article approached this question along a certain line of inquiry to reach its conclusions. Hopefully, at the least it will be helpful in furthering the dialog on this topic. This article set out to compare a passively managed GIC portfolio, a surrogate GIC index, against public market fixed income indices to demonstrate the risk/return efficiency of GICs using the familiar framework of portfolio indices. GICs seem to compare efficiently with traditional fixed income securities of similar duration and credit quality. This is true in spite of the dilemma that currently GICs exist in their own sector outside of the benchmark composites. As for liquidity risk, the GIC return series' were created out of GIC rates that were net of benefit responsive charges. In a sense, this corrects for the fact that these are private bonds being compared against public bond markets. GICs with benefit responsive insurance have very efficient liquidity built into their structure and covenants. While this article didn't focus on active management within GICs, it is worth noting that the frequent change in sign of the excess returns from month to month between GICs and the public markets suggest that many tactical opportunities exist between these two markets. In addition, with the absence of an economic returns-based GIC index, this analysis also demonstrates the efficacy of using a public bond index to benchmark a GIC portfolio. The challenge will be for a stable value manager to fit an appropriate index to the GIC strategy that he or she is utilizing. Oftentimes, an allocation to GICs is mandated by strategic policy for inclusion in these funds. This analysis points to a conclusion that these allocations should not be removed from the discretionary holdings of a stable value manger for the purposes of compiling a performance composite.
Read Next: DC Plan Brokerage Accounts:
|
|
