Statistical validation of financial time series via visibility graph

TL;DR

This paper introduces a statistical validation method for visibility graphs derived from financial time series, providing a new market indicator that correlates with and predicts financial instability periods.

Contribution

It develops a validation procedure for visibility graph links using ARCH model-based null hypotheses and proposes a novel market indicator linked to financial stability.

Findings

Validated visibility graph links against ARCH null hypothesis

Proposed a market indicator correlated with financial instability

Indicator predicts periods of financial crises

Abstract

Statistical physics of complex systems exploits network theory not only to model, but also to effectively extract information from many dynamical real-world systems. A pivotal case of study is given by financial systems: market prediction represents an unsolved scientific challenge yet with crucial implications for society, as financial crises have devastating effects on real economies. Thus, nowadays the quest for a robust estimator of market efficiency is both a scientific and institutional priority. In this work we study the visibility graphs built from the time series of several trade market indices. We propose a validation procedure for each link of these graphs against a null hypothesis derived from ARCH-type modeling of such series. Building on this framework, we devise a market indicator that turns out to be highly correlated and even predictive of financial instability periods.