Revealing high-frequency trading provision of liquidity with visualization

Liquidity is crucial for successful financial markets. It ensures that all investors are able to buy and sell assets quickly at a fair price. High Frequency Traders (HFTs) utilize sophisticated algorithms operating with extreme speed and are frequently cited as liquidity providers. The objective of this paper is to investigate the liquidity provision of a number of HFTs to determine their effects on aggregate marketplace liquidity. We consider a large data set collected from the Australian Securities Exchange throughout 2013, providing a near complete picture of all trading activity. Our method is to consider temporal bar charts, association scatterplots, faceted plots and network diagrams to provide visualizations that yield both novel and conventional insights into how HFTs are operating in the market, specifically with respect to liquidity provision. Consistent with HFTs avoiding adverse selection, our results show that on aggregate, HFTs often consume rather than provide liquidity. Furthermore, liquidity consumption often occurs very quickly over intra-millisecond time periods. We conclude that HFTs are not exclusively focused on liquidity provision.