Visualizing the Trading Networks: An Analysis of Systemic Risk in High Frequency Trading Markets

Abstract

As stock exchanges replaced physical trading floors with an electronic environment, high frequency trading began to emerge. While high frequency traders theoretically provide liquidity, certain behaviours have been criticized for increasing instability in the market. Some of high frequency trading’s most infamous strategies include spoofing, where brokers cancel trades within milliseconds of submission; and quote stuffing, where traders overwhelm the exchanges with large number of orders, cancellations, and updates to slow down the system and hide their own strategy. Brokers may also put in orders with prices far from the bid and ask price, meaning that they do not actually intend to have the orders executed. Because most of these actions occur anonymously, it is difficult to determine the source of these orders. While most order books do not reveal the identity of the buyer and seller of each trade, the Toronto Stock Exchange does. With this new information, this study investigates how high frequency traders interact with each other by plotting an animation the trading network over time. In addition, this paper will identify the key players of the Toronto Stock Exchange and will perform a theoretical exercise that removes these active traders from the network. In doing so, this paper will study how the trading networks react to the missing broker. The resulting state of the network provides great implications on the amount of influence a single, active broker may have on the trading network, which in turn indicates the level of systemic risk in high frequency trading markets.