Analysis of extreme dependence between Istanbul Stock Exchange and Brent Oil Returns using bivariate extreme value theory

Abstract

In this thesis, the relationship between oil price movements and Turkish stock market is investigated. Given the fact that Turkey is an emerging and oil dependent country, the relationship between stock market behavior and fluctuations in oil prices is analyzed by focusing on extreme observations. Bivariate extreme value methodology is used in order to examine the dependence structure between oil and stock market (ISE100). The residuals of autoregressive integrated moving average (ARIMA) models of stock market index and Brent oil returns are examined by using bivariate extreme value analysis over the period between 1988 and 2012. The overall period studied is subdivided into two phases for analysis. Results of the thesis indicate a higher dependence in the second phase (2000-2012), compared to the first phase (1988-2000). Results also show that in the second phase the extremes on the negative tails coincide more commonly compared to the extremes on the positive tails, which is in line with the current literature findings. A more focused analysis of bivariate extreme value theory (EVT) for the 2008-2012 period is conducted within the thesis and highest dependence of oil prices and stock market is found in this period. However, in general Turkish stock market and oil returns are asymptotically independent in extreme observations, which suggest different diversification strategies for portfolio managers.