Quantifying the randomness of the forex market

Currency markets are international networks of participants opened all day during weekdays without a supervisory entity. The precise value of an exchange pair is determined by the decisions of the central banks and the behavior of the speculators, whose actions can be determined on the spot or be related to previous decisions. All those decisions affect the complexity and predictability of the system, which are quantitatively analyzed in this paper. For this purpose, we compare the randomness of the most traded currencies in the forex market using the Pincus Index. We extend the development of this methodology to include multidimensionality in the embedding dimension, to capture the influence of the past in current decisions and to analyze different frequencies within the data with a multiscale approach. We show that, in general, the forex market is more predictable using one hour ticks than using daily data for the six major pairs, and present evidence suggesting that the variance is easier to predict for longer time frames. (C) 2021 Elsevier B.V. All rights reserved.

Automated summary

Currency markets are international networks open all day without supervision, where the value is determined by central bank decisions and speculator behavior. This study quantitatively analyzes the impact of decisions on system complexity and predictability, showing that forex market is more predictable on an hourly scale and variance is easier to predict for longer time frames.