Testing the covariance stationarity of CEE stocks

This paper investigates whether the daily stock returns of the Polish, Czech and Hungarian stock markets are covariance stationary. Using the Pagan – Schwert (1990) and Loretan – Phillips (1994) testing procedures, we show that contrary to the widely accepted assumption of covariance stationarity, the stock returns in Central and Eastern European (CEE) stock markets do not appear to be covariance stationary. Our results further suggest that the occurrence of unconditional volatility shifts appears to be synchronized across stocks.