How Often to Sample a Continuous-Time Process in the Presence of Market Microstructure Noise
Top Cited Papers
- 10 February 2005
- journal article
- research article
- Published by Oxford University Press (OUP) in The Review of Financial Studies
- Vol. 18 (2) , 351-416
- https://doi.org/10.1093/rfs/hhi016
Abstract
In theory, the sum of squares of log returns sampled at high frequency estimates their variance. When market microstructure noise is present but unaccounted for, however, we show that the optimal sampling frequency is finite and derives its closed-form expression. But even with optimal sampling, using say 5-min returns when transactions are recorded every second, a vast amount of data is discarded, in contradiction to basic statistical principles. We demonstrate that modeling the noise and using all the data is a better solution, even if one misspecifies the noise distribution. So the answer is: sample as often as possible.Keywords
This publication has 38 references indexed in Scilit:
- Modeling and Forecasting Realized VolatilityEconometrica, 2003
- The Effects of Random and Discrete Sampling when Estimating Continuous-Time DiffusionsEconometrica, 2003
- Fully Nonparametric Estimation of Scalar Diffusion ModelsEconometrica, 2003
- Econometric Analysis of Realized Volatility and its Use in Estimating Stochastic Volatility ModelsJournal of the Royal Statistical Society Series B: Statistical Methodology, 2002
- Maximum Likelihood Estimation of Discretely Sampled Diffusions: A Closed-form Approximation ApproachEconometrica, 2002
- The Distribution of Realized Exchange Rate VolatilityJournal of the American Statistical Association, 2001
- Nonparametric Pricing of Interest Rate Derivative SecuritiesEconometrica, 1996
- Bid-ask spreads in the interbank foreign exchange marketsJournal of Financial Economics, 1994
- On the Estimation of Bid-Ask Spreads: Theory and EvidenceJournal of Financial and Quantitative Analysis, 1988
- NoiseThe Journal of Finance, 1986