Herding Behaviour, Bubbles and Log Periodic Power Laws in Illiquid Stock Markets. A Case Study on the Bucharest Stock Exchange

In this paper we investigate the herding behaviour of the Bucharest Stock Exchange (BSE), using log periodic power laws models. By analysing the behaviour of the most speculative index from the Bucharest Stock Exchange, the BET-FI, we are able to demonstrate that Log-Periodic Power Law (LPPL) models are a useful tool for recognizing the behaviour of a stock market bubble, and have good abilities for predicting the critical point of a bubble. From our statistical investigation, it turns out that an iterative calibration of the model for the BET-FI regime leads ex post to a rather accurate forecast of the stock market crash in January 2008. Next, by using the same methodology, the anti-bubble regime from 2008 is used for a statistical fit. We then find an accurate “prediction” of the local point of phase transition on 27 October 2008.