Quantifying reflexivity in financial markets: towards a prediction of flash crashes

TL;DR

This paper introduces a measure of market endogeneity using Hawkes models to quantify endogenous feedback in financial markets, revealing a significant increase in internal activity from 1998 to 2010 and providing a way to assess market stability.

Contribution

It presents a novel quantitative measure of market reflexivity based on Hawkes processes, linking endogenous activity to market stability and flash crash prediction.

Findings

Endogeneity increased from 70% in 1998 to less than 30% since 2007.

The measure quantifies how close markets are to a critical state of instability.

Market activity can be monitored to assess risk of flash crashes.

Abstract

We introduce a new measure of activity of financial markets that provides a direct access to their level of endogeneity. This measure quantifies how much of price changes are due to endogenous feedback processes, as opposed to exogenous news. For this, we calibrate the self-excited conditional Poisson Hawkes model, which combines in a natural and parsimonious way exogenous influences with self-excited dynamics, to the E-mini S&P 500 futures contracts traded in the Chicago Mercantile Exchange from 1998 to 2010. We find that the level of endogeneity has increased significantly from 1998 to 2010, with only 70% in 1998 to less than 30% since 2007 of the price changes resulting from some revealed exogenous information. Analogous to nuclear plant safety concerned with avoiding "criticality", our measure provides a direct quantification of the distance of the financial market to a critical state defined precisely as the limit of diverging trading activity in absence of any external driving.