Application of the generalized method of moments for estimating continuous-time models of U.S. short-term interest rates

We show by Monte Carlo simulations that the jackknife estimation of QUENOUILLE (1956) provides substantial bias reduction for the estimation of short-term interest rate models applied in CHAN ET AL. (1992) - hereafter CKLS (1992). We find that an alternative estimation based on NOWMAN (1997) does not sufficiently solve the problem of time aggregation. We provide empirical distributions for parameter tests depending on the elasticity of conditional variance. Using three-month U.S. Treasury bill yields and the Federal fund rates, we demonstrate that the estimation results can depend on both the sampling frequency and the proxy that is used for interest rates.