IKTISAT ISLETME VE FINANS, sa.296, ss.77-91, 2010 (SSCI)
This paper investigates whether the Istanbul Stock Exchange (ISE) prices can be characterized as a random walk or mean reversion process in a non-linear framework. We employ an unrestricted two-regime threshold autoregressive (TAR) model with an autoregressive unit root based on bootstrap methods. Using weekly observations, our findings indicate that weekly ISE-100 data is non-linear and characterized by a unit root process over the period January 1999 to January 2009. To test the presence of non-linear dependence, we apply the BDS test to the residuals of estimated TAR model and reject the null hypothesis of independently and identically distributed (IID) assumption. The main result is that the paper rejects the random walk hypothesis for the Istanbul Stock Exchange.