Exit strategies in bubble-like markets using a changepoint model

Abstract

We present applications of a stochastic changepoint detection model in the context of bubble-like financial markets. A changepoint of a random sequence is an unknown moment of time when its trend changes. The aim is to detect a direction change in a sequence of stock market or other asset index values, while sequentially observing it. A detection rule thus models an exit strategy before a possible market crash. We describe theoretical results and apply them to several stock market bubbles including stock markets in the US in 1929, 1987, 2008, and China in 2015.