High-Frequency Trading: How Money Flow Affects Stock Returns

Abstract: Purpose:
The main purpose with this thesis is to study tick-data in order to see if intra-day volume can predict short-term market movements.

Methodology:
This thesis a quantitative study, using a deductive method and with an exploratory research design.

Theoretical framework:
In the theoretical framework, important theories for this thesis are presented, such as tape reading, high-frequency trading, and previous findings about money flow and measures of floor trading.

Empirical foundations:
The data sample consists of approximately 2.2 million trades during January 1, 2005 to April 22, 2010. The data was retrieved from Six Telekurs database.

Conclusions:
The high-frequency measures used in this thesis showed a significant relationship between the measures and the stock return at a 1% significance level. The accumulated money flow was highly positively correlated with the stock return until late 2008, and since then it became negatively correlated. Steadily increasing activity by high-frequency trading algorithms, as well as the tick-size changes that occurred during 2009, might be an explanation. The trading model, which used the three measures studied, had almost twice as high risk-adjusted return as the stock itself.