Abstract:

In this study, we present new evidence with the Koenker-Xiao (2004) quantile unit root test based on the quantile autoregression approach, which provides robust inferences for non-normal processes, to examine the asymmetric dynamic process of the BIST100 Borsa Istanbul index. The quantile autoregression approach allows to measure the persistence of shocks of different magnitudes and signals affecting the stock markets. This approach can catch the adjustment of asymmetrical dynamics in the long-run equilibrium of the index. Therefore, quantile unit root tests add new approaches to stock market dynamics compared to traditional unit root methodologies based on least squares regression method. Our results, although not consistent at all frequencies, show that market behavior not only returns to the mean, but also behaves differently to positive and negative shocks. Compared to traditional unit root tests, quantile unit root tests provide more evidence to support the efficiency of the Turkish stock market. We also show that asymmetric inferences in shock magnitude and sign are helpful in detecting deviations from information efficiency in terms of asset pricing in stock markets.

Keywords: