haku: @all frequency / yhteensä: 698
viite: 15 / 698
| Tekijä: | Djupsjöbacka, D. |
| Otsikko: | Implications of market microstructure for realized variance measurement |
| Lehti: | European Journal of Finance
2010 : JAN/FEB, VOL. 16:1-2, p. 27-43 |
| Asiasana: | markets volatility measurement monte carlo technique simulation autocorrelation sampling |
| Kieli: | eng |
| Tiivistelmä: | During the last years, volatility (henceforth as: vol-ty) measuring and estimation based on intra-day high-frequency data has become more popular. Based on the theory of quadratic variations (here as: var/s.) of diffusions, it can be shown that realized var/s. measures, based on sufficiently frequently sampled returns, are error-free vol-ty estimates. This study aims at examining realized var/s. measures, where well-documented market microstructure effects, e.g. return autocorrelation and vol-ty clustering, are included in the return generating process. It is found that the use of squared returns as a measure for realized var/s. will lead to estimation errors on sampling frequencies adopted in the literature. In the case of return autocorrelation, there will be systematic biases. Furthermore, there is increased standard deviation in the error btw. measured and real var/s. with sampling frequency decreasing and with non-constant vol-ty. |
SCIMA