High frequency price impact of order book events - Evidence from Nordic equity markets

The thesis provides new empirical evidence on the limit order book activity in NASDAQ OMX Nordic and studies the price impact of order book events in short time intervals. The objective is to evaluate if the order flow imbalance is a stronger driver behind asset price movements than the directional imbalance of executed trades or the total volume. DATA AND METHODOLOGY The high frequency data set is OMX NASDAQ TotalView-ITCH feed that includes full order book activity information in March 2010 and September 2012 in NASDAQ OMX Nordic. The data set includes 43 trading days. 50 most traded stocks are selected for the study covering NASDAQ OMX Helsinki, Stockholm and Copenhagen. Regression analysis is used to analyze the mid-quote price impact of the order book events during a ten second time interval based on the order flow imbalance. In addition, the study covers intraday patterns in the asset price dynamics and applies the order flow based measure to explain changes in trade prices. RESULTS Nordic equity markets are heavily driven by high frequency trading and intensive quote submission activity. Only 5.7% of the submitted visible limit orders are filled in full or partially in September 2012. Corresponding figure for March 2010 is 8.0%. The average number of daily submitted limit orders per stock increased 68% to 60 thousand orders between the two time periods. Majority of the limit orders are cancelled in less than ten seconds after the submission. Among the most actively traded stocks, the cancellations take place within the first two seconds. The study supports earlier results that high frequency price changes are strongly driven by the order flow imbalance. The result is robust to changes in time period, selected stocks and intraday seasonality. The order flow imbalance is sufficient to explain on average 57% of the mid-quote price impact. Price impact is empirically observed to approximately follow a concave square root relation with the total traded volume. The price impact of order book events is highest in the morning and decreases during the trading day. The order flow imbalance is found to be a better measure for the price impact than the trade imbalance or the total volume, and it is shown to have statistically significant effect on explaining the trade price changes. In addition, the study finds considerable improvement in the goodness of fit of the model in September 2012 compared to March 2010. Between the two periods the order to trade ratio increased from 11.1 to 15.9. Order flow is becoming more significant as the driver in asset price movements as the dynamics in the microstructure environment continue to evolve