Identifying Bull and Bear Market Regimes With a Robust Rule-Based Method

Abstract

A new method for identifying bull and bear financial market regimes is proposed, related to a classic algorithm for picking turning points in the business cycle. Our approach uses only a single discrete parameter, adjusted to the periodicity of the data, which largely removes subjectivity from the regime identification process and limits data snooping. Applying it to the Dow Jones Industrial Average index data, we show its capability of obtaining a classification similar to competing multi-parameter methods, without imposing any conditions on regime duration or amplitude. Our algorithm can be easily applied across different asset classes, where its direct competitors may fail, as we show in an out-of-sample identification example for oil price series and an exchange rate.

Keywords: Financial market, Market regimes, Financial cycle, Bull market, Bear market

JEL Classification: G10, C18