A Method for Detecting Structural Breaks and an Application to the Turkish Stock Market

We suggest a procedure for model update, based on detection of structural breaks at unknown change-points. The procedure makes use of the SupF test introduced by Andrews (1993). We apply this procedure for modelling the common stock index returns in the Istanbul Stock Exchange for the 11 year period of 1989 - 1999. The underlying model consists simply of a mean plus noise, with occasional jumps in the level of mean at unknown time instances. The problem is the detection of this jump and the corresponding model update. We find critical values for the SupF test statistic by using the Bootstrap method. A trading rule that uses the forecasts from the suggested procedure is observed to outperform the buy-and-hold strategy