Although financial markets have become more and more complex, they have still managed to retain a degree of predictability. One predictable phenomenon of is how the trading of financial assets can be modeled by a model that exhibits cascading. Cascading behavior is the idea of one event triggering the occurrence of a subsequent event. In the financial markets this can be thought of people mocking the same trade as someone.

It has been shown that significant trades in the market cause subsequent trades of the same sign (buy/sell), and thus increase the arrival rates at that specific time which eventually decays over time.

This behavior is at the heart of this study and we model the arrival rates of the following:

  • Large Day Trades 
  • Small Day Trades
  • Large Night Trades
  • Small Night Trades

Daytime is defined as 7AM CST – 3:15 CST while Nighttime is defined as 5PM – CST – 7AM CST. We split up these time periods when modeling because trading activity and behavior is different between both time periods – daytime is much more active.

Furthermore, we also split up trades by “size” into Large and Small for two reasons:

(1) Significantly more small trades occur during both day and night, and thus their arrival intensities would be different.

(2) Large trades are generally given less favorable prices and thus cause larger movements in price. Knowing their arrival rates could help one to make inferences about price volatility.

By being able to predict how quickly trades arrive, we can develop intuition about trading behavior and perhaps implement a trading strategy to profit from.