Detecting Structural Breaks in Cryptocurrency Market

Abstract

This paper aims to compare the empirical performance of two approaches in detecting structural breaks and outliers due to the significant frequent price changes seen in cryptocurrencies.The two approaches are indicator saturation (IS) and Bai and Perron (BP).The cryptocurrency data employed in this study are Bitcoin and Ethereum.In comparing the performance of the two approaches, this study performed multiple empirical comparisons using various significant levels, different data frequencies, as well as the original and log series (price).The findings showed that the prices contained structural breaks and outliers and that the IS approach performed significantly better than the BP test in terms of the identified structural breaks as well as outliers across different settings.The contribution of this study is providing empirical comparisons between IS and BP approaches using cryptocurrency data.These findings are important to the potential stakeholders, in particular, for quality control in industries, for setting price targets, and for confirming trading signals to reduce potential losses.