Score-Driven GARCH for Irregular, Discrete Intraday Prices

Summary

The document presents a high-frequency price model designed for irregular observation times and discrete price changes. It uses a zero-inflated Skellam distribution within a score-driven volatility framework, with a moving-average component intended to filter market microstructure noise. Smoothing splines represent how trade durations relate to volatility and how both vary throughout the trading day.

The parameters are estimated by maximum likelihood. An empirical application to IBM stock is reported to fit the data well, and the authors note that the model can also measure daily realized volatility. The account offers no detailed comparison with alternative models, quantitative fit statistics, or broader tests across assets and market conditions, so its evidence is limited to the stated single-stock study.

Key ideas

• The model accommodates irregularly timed observations and discrete price changes.
• A zero-inflated Skellam distribution drives a score-based, time-varying volatility process.
• A moving-average component is used to account for market microstructure noise.
• Smoothing splines capture links between trade duration and volatility, including intraday patterns.
• The approach is estimated by maximum likelihood and is also proposed for daily realized volatility measurement.

Tags

# An Intraday GARCH Model for Discrete Price Changes and Irregularly Spaced Observations


# An Intraday GARCH Model for Discrete Price Changes and Irregularly Spaced Observations









We develop a novel observation-driven model for high-frequency prices. We account for irregularly spaced observations, simultaneous transactions, discreteness of prices, and market microstructure noise. The relation between trade durations and price volatility, as well as intraday patterns of trade durations and price volatility, is captured using smoothing splines. The dynamic model is based on the zero-inflated Skellam distribution with time-varying volatility in a score-driven framework. Market microstructure noise is filtered by including a moving average component. The model is estimated by the maximum likelihood method. In an empirical study of the IBM stock, we demonstrate that the model provides a good fit to the data. Besides modeling intraday volatility, it can also be used to measure daily realized volatility.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.