Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

TailCoR: A new and simple metric for tail correlations that disentangles the linear and nonlinear dependencies that cause extreme co-movements

PLoS One. 2023 Jan 3;18(1):e0278599. doi: 10.1371/journal.pone.0278599. eCollection 2023.

Abstract

Economic and financial crises are characterised by unusually large events. These tail events co-move because of linear and/or nonlinear dependencies. We introduce TailCoR, a metric that combines (and disentangles) these linear and non-linear dependencies. TailCoR between two variables is based on the tail inter quantile range of a simple projection. It is dimension-free, and, unlike competing metrics, it performs well in small samples and no optimisations are needed. Indeed, TailCoR requires a few lines of coding and it is very fast. A Monte Carlo analysis confirms the goodness of the metric, which is illustrated on a sample of 21 daily financial market indexes across the globe and for 20 years. The estimated TailCoRs are in line with the financial and economic events, such as the 2008 great financial crisis and the 2020 pandemic.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Monte Carlo Method*

Grants and funding

SB was supported by a grant (165880) as a PhD Fellow of the Research Foundation- Flanders (FWO, https://www.fwo.be). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.