Evolutionary correlation, regime switching, spectral dynamics and optimal trading strategies for cryptocurrencies and equities

TL;DR

This paper compares the evolutionary dynamics of cryptocurrencies and equities using advanced spectral and correlation analysis, revealing key differences and informing optimal trading strategies for both asset classes.

Contribution

It introduces a comprehensive methodology combining random matrix theory, PCA, and spectral analysis to study structural breaks and sector dynamics in cryptocurrencies and equities.

Findings

Distinct eigenspectra for cryptocurrencies and equities

Time-varying sector behaviors identified in both markets

Portfolio algorithms highlight different strategies for crypto and stocks

Abstract

This paper uses new and recently established methodologies to study the evolutionary dynamics of the cryptocurrency market, and compares the findings with that of the equity market. We begin by applying random matrix theory and principal components analysis (PCA) to correlation matrices of both collections, highlighting clear differences in the eigenspectra exhibited. We then explore the heterogeneity of both asset classes, studying the time-varying dynamics of underlying sector behaviours, and determine the collective similarity within each collection. We then turn to a study of structural break dynamics and evolutionary power spectra, where we quantify the collective affinity in structural breaks and evolutionary behaviours of underlying sector time series. Finally, we implement two algorithms simulating `portfolio choice' dynamics to compare the effectiveness of stock selection and sector allocation in cryptocurrency portfolios. There, we highlight the importance of both endeavours and comment on noteworthy implications for cryptocurrency portfolio management.