H.Wendt , P. Abry, Y. Malevergne, M. Senneret, G. Perrin, S. Jaffard
In this work, we aim to study the characteristics of the joint time dynamics of a basket of Foreign Exchange prices.
Building on recent theoretical developments enabling multivariate multifractal analysis of short time series using Bayesian formalism, we carry out a longitudinal analysis of six series of Foreign Exchange Rates over a 17-year period, in 6-month rolling steps.
The analyses confirm the market efficiency hypothesis (absence of temporal correlation), but also highlight a joint dependency structure of these six series in terms of joint occurrences of local statistical structures, which we can interpret as a joint inter-correlation of the volatilities associated with these rates.
This suggests that the ForEx market is mainly governed by a single clock around the world.
This paper was published in EUSIPCO 2024, 32nd European Signal Processing Conference.
> Download the research article
> Download presentation (slides)