An evaluation of alternative methods used in the estimation of Gaussian term structure models

The mean–variance (in)efficiency of duration‐based immunization

“…On the other hand, if one restricts the analysis to nonoverlapping immunization strategies, the time series is too short to obtain a sufficient number of observations.14 The data can be downloaded from: www.federalreserve.gov/pubs/feds/2006/200628/200628abs.html. Daily timeseries for the six parameters of the Nelson-Siegel-Svensson yield curve fit are available (seeGürkaynak et al, 2007, for the detailed methodology).15 By means of simulations,Juneja (2015) provides evidence that application of the Kalman filter to Gaussian affine term structure models yields an accurate parameter estimation even in the presence of serial and cross-sectional correlations in the error term.16 The estimation using weekly data returned very similar parameter values.17 Intuitively, if w is too high, the generated yield curves will be too close to the original yield curves (at the limit, they would be mere translations across time of the actual yield curves). If w is too low, the generated yield curves will lose most of the statistical properties of the original yield curves.…”

“…15 15 By means of simulations, Juneja (2015) provides evidence that application of the Kalman filter to Gaussian affine term structure models yields an accurate parameter estimation even in the presence of serial and cross‐sectional correlations in the error term. …”