Abstract
This study examines the performance of market order execution strategy in a pure limit order driven environment based on three bid-ask spread forecasting models. While a naive spread forecasting model based on previous day's spread and average 10 trading days’ spread could deliver a cost saving of 3.94% and 14.87%, this benefit increases to 22.14% for a seasonal autoregressive moving average spread forecasting model. The empirical results are evident that the intraday spread forecasting benefits follow a downward-sloping pattern.
Notes
1 Due to space limitation, the results of Diebold and Mariano's (1995) test are not presented but are available upon request.