THE BOND MARKETS: CONCLUDING COMMENTS 2

It is worthwhile suggesting further avenues of research, which would benefit from the framework of this paper. First, the issue of what information releases cause jumps is an open question. Locating jumps in the data and associating them with market events is one way of addressing this question. Second, a question of importance is whether Fed actions are endogenous or exogenous to the interest rate markets. This aspect is a strong determinant in the choice of the modelling framework (see Balduzzi, Bertola and Foresi [8]).

Third, the shape of the term structure is also a function of whether agents price jump risk or not. Using the pricing equations in this study, cross sectional estimation of parameters would shed light on this question. More info This would also address the issue of whether the expectations hypothesis (see CIR [21]) holds in a Poisson-Gaussian world.

Fourth, rather than model jumps in the level of the interest rate, modelling jumps in the mean and volatility of the short rate is an alternate approach (see Naik k, Lee [37]). Fifth, in a recent innovation, Heston [30] employs a gamma process as an alternative to the Poisson-Gaussian framework. A comparison of this model with the one in this paper would be insightful. Sixth, this work may be related to the work of Brandt and Santa-Clara [13], who develop a method of estimation using simulated transition density functions.

Their work relates to diffusions only, and hence may be extended to jump-diffusions and then confirmed using the closed-form results of this paper. Finally, examining very short frequency intra-day data may reveal better the possible causes of jumps in bond yields. Eurodollar yields have also been suggested as a better benchmark for tests of this sort (sec Duffee [27]). This paper leaves this rich menu of research projects for future work.
w6631-27
w6631-28