| Tekijä: | Sandmann, G. Koopman, S. J. |
| Otsikko: | Estimation of stochastic volatility models via Monte Carlo maximum likelihood |
| Lehti: | Journal of Econometrics
1998 : DEC, VOL. 87:2, p. 271-301 |
| Asiasana: | ECONOMETRICS ESTIMATION STOCHASTIC PROCESSES VOLATILITY MONTE CARLO TECHNIQUE |
| Vapaa asiasana: | GARCH MODEL IMPORTANCE SAMPLING KALMAN FILTER SMOOTHER |
| Kieli: | eng |
| Tiivistelmä: | The Monte Carlo maximum likelihood method (MCL) of estimating stochastic volatility (SV) models, is discussed in this article. The representation of the model is in a linear state space form so the Kalmar filter can be employed to compute the Gaussian likelihood function via the prediction error decomposition. Due to the log chi-square disturbances in the measurement equation of the model, the Gaussian likelihood will only make up a part of the true likelihood function. The proposed MCL estimator approximates the remainder term via Monte Carlo simulation. The finite sample performance of the MCL is examined in a simulation experiment. |
SCIMA