The Benefits of Resiliency to Standard Market Impact Models

Abstract

We leverage Kepler Cheuvreux client order database over the period October 2014 — October 2016 (349,442 trades corresponding to a EUR92.3bn turnover) to estimate new models of market impact. We find a multiplicative relationship between the market impact and the explanatory factors (the volatility, the trading period participation rate and the trading duration). Furthermore, the relationship between the participation rate and the duration on one side and the market impact on the other is concave, with the effect of the participation rate on the market impact scaling as a square root. We introduce a new indicator of resiliency, which measures the ability of the order book to resist the aggressive order flow in a given period. This indicator shows a positive correlation with the residuals of our standard model of market impact, clearly demonstrating that the more resilient the stock, the more resistant it is to market impact. Thus, we are able to calibrate an enhanced model of market impact using our indicator of resiliency, which improves the explanatory power of the model compared to standard approaches. Our resiliency indicator thereby exposes the relationship between the market impact at the meta-order scale and the market impact at the elementary trade scale.