Title:Dissecting cross-impact on stock markets: An empirical analysis

Abstract:The vast majority of recent studies in market impact assess each product individually, and the interactions between their order flows are disregarded. This strong approximation may lead to an underestimation of trading costs and possible contagion effects. Transactions mediate a significant part of the interaction between different instruments. In turn, liquidity shares the sectorial structure of market correlations, which can be encoded as a set of eigenvalues and eigenvectors. We introduce a multivariate linear propagator model that successfully describes such a structure, and reproduces well the response and a significant fraction of the covariance matrix of returns. We explain in detail the various dynamical mechanisms that contribute to these quantities. We also define two simplified models with substantially less parameters to reduce overfitting, and show that they have superior out-of-sample performance.