|
Abstract:
|
Volatility estimation is a topic of much interest to market participants. In particular, changes observed in agricultural commodity markets as a result of the switch from pit to electronic trading have motivated new research that incorporates the area of market microstructure into traditional volatility estimators. On one hand, some studies suggest that using high-frequency data improves volatility measures because of its rich and frequently sampled nature. However, other studies argue that microstructure effects such as the bid-ask spread and discreteness of observations have adverse effects, and therefore high-frequency data does not necessarily bring improvements to volatility estimation. In addition, most models being used do not take into account the information contained in the quotes and depths at different levels of the limit order book (LOB). We assess volatility models which use data at different frequencies (i.e., two time scales estimator) and incorporate the information contained in the LOB (i.e., GARCH) in agricultural commodity markets. The results should be important for the development of risk management strategies.
|
