Nonlinearity in Stock Exchange Markets: The Case of Bist 100 Indices

Abstract

In this paper, using data for the Bist 100 index, we investigate the presence of nonlinearities by employing several nonlinearity tests. The Brock, Dechert, and Scheinkman (BDS) and runs tests were first applied to the series to show an initial indication of nonlinearity. The findings for the BDS and runs test of randomness were followed by other sets of direct nonlinearity tests developed by White (1989), Terasvirta (1993), Keenan (1985), and Tsay (1986). Also, the Threshold Autoregression (TAR) test is employed as a final test to confirm the existence of nonlinearity in the Turkish stock exchange market. From the results of the nonlinearity test, it is concluded that the Bist 100 index is characterised by the presence of nonlinearities and cycles. This finding is in contrast with the efficient market hypothesis (EMH) implying that the Turkish stock exchange market is inefficient.