Strip bonds and the efficiency of the bond market
A recent study (Government of Canada Yield-Curve Dynamics, 1986-2003, by Grahame Johnson) published in the Bank of Canada Review concluded that the Canadian strip bond service, among other factors, contributed to making the Canadian bond market safer, or less risky, for investors, during the during the 1990's, as bonds are now priced more consistently based on the yield curve than before. The study includes extensive statistical evidence in support of this.
The study attributes this to several factors, specifically citing changes made by CDS such as 1) the introduction of generic CUSIPs for strip bonds, strip bond packaging, and reconstitution in 1993, and 2) the changes in 2001 that allowed bonds to be reconstituted beyond the amount originally stripped. Other improvements included a benchmark program, formal consultations with market participants, a bond buy-back program, and measures to improve the integrity of the market.
Bond traders in Canada have told me for several years now that the CDS strip bond service enhancements forced bond market participants to value bonds based on the present value of cash flows, discounted using current yields. Prior to these changes, it was quite normal to see a wide difference in yields, usually 0.60% to 0.70% or more, on some Government of Canada bonds having the same maturity dates.
In early 1993, one money manager from a well-known U.S. pension fund told me that he could not understand why Canadians were undervaluing these bonds so much when differences of this magnitude did not exist in the U.S. The undervaluation made it clear that many investors were avoiding certain Government of Canada bonds. The yield differences narrowed substantially in 1993, when CDS implemented generic CUSIPs for strip bonds, reconstitution, and packaging. In 2001, the variance was virtually eliminated when CDS began allowing reconstitution beyond the amount originally stripped.
Today, if you check the papers, the yields for bonds with the same maturity dates are either the same or within 0.01%. Interestingly, in the U.S., where the strip bond service is not as advanced as in Canada, larger yield differences still exist for government bonds having the same maturity dates, even though they implemented many of the other changes that the authors believe made the bond market less risky (a bond buy-back program, benchmark bonds, formal consultations, etc.).
A more technical and lengthier analysis was published as a Working Paper (#2004-48: An Empirical Analysis of the Canadian Term Structure of Zero-Coupon Interest Rates by David J. Bolder, Grahame Johnson, and Adam Metzler).
Copyright Keith Campbell ©2005. All rights reserved.

Next