Volatility of Stock Market Indices - An Analysis based on SEMIFAR Models

By applying SEMIFAR models (Beran, 1999), we examine 'long memory' in the volatility of worldwide stock market indices. Our analysis yields strong evidence of 'long memory' in stock market volatility, either in terms of stochastic long-range dependence or in form of deterministic trends. In some cases, both components are detected in the data. Thus, at least partially, there appears to be even stronger and more systematic 'long memory', than suggested by a stationary model with long-range dependence.