Sudden Changes in Volatility - the Case of the Five Financial Investment Companies in Romania

This paper studies the sudden changes in volatility of the five most traded shares on the Bucharest Stock Exchange Financial Investment Companies, by using the ICSS algorithm proposed by Inclan and Tiao (1994). Events leading to unexpected changes in variance are predominantly local ones; the only significant global event, with negative influence on the volatility regime is the evolution of foreign markets in 2008-2009, following the global financial crisis. In terms of persistence in volatility, it is found that the false long memory effect is gone when the dummy variables associated to events that have caused sudden changes in volatility are incorporated in the GARCH model.