On Portfolio Theory, Holding Period Assumptions, and Bond Maturity Diversification

Abstract

* This paper is as much a warning for readers of holding period-oriented articles in general as it is a commentary on the work of Jess B. Yawitz, George H. Hempel, and William J. Marshall [11]. What follows is a general criticism of the blind application of specific holding periods, showing how the choice of a holding period can reduce or invalidate the tenets of modern portfolio theory. Several important recent papers apply MarkowitzSharpe models to bond portfolio selection only to develop limited or negative findings. The general results of such studies [6, 7, 11, 12] assign not more than a limited value to maturity diversification, contrary to the implications of formal portfolio theory. Among these, the work of Yawitz, Hempel, and Marshall [11] provides a useful but limited practical application of portfolio theory to the selection of government bonds. Although we elaborate primarily on this paper, we will also briefly discuss other relevant research that warrants similar warnings. The Yawitz et al. treatment of benefits from maturity diversification unnecessarily restricts the scope of their model. Our discussion complements their work, adding two additional conclusions on practical management of government bond portfolios. First, we reconcile their admittedly arbitrary choice of holding period lengths with recent research on how the holding period chosen may affect the efficient portfolio selected [3]; second, we show that opportunities to pool risk in the government bond market via diversification by maturities may be less limited than Yawitz, Hempel, and Marshall suggest. Additionally, Gressis, Philippatos, and Hayya [3] demonstrate that the mean-variance efficient frontier