Investigation of equity return volatilities with nonlinear methods and investigation of market efficiency: A comparative analysis with BRICS-T countries

The aim of this dissertation is to examine the volatility in the stock markets of BRICS-T countries with nonlinear time series methods and to investigate market efficiency. In the study, the returns of indices at daily frequency basis, covering the dates between 01.01.2011 and 01.01.2021, were analyzed by applying logarithmic difference transformation. First of all, descriptive statistics of the series consisting of mean, standard deviation, skewness, kurtosis, Jarque-Bera test and ARCH-LM tests were examined. Then, unit root analysis was carried out with ADF, PP and KPSS tests. Although there are findings on the existence of the efficient markets hypothesis, it has been determined that the stock markets of the countries examined are not efficient. The linearity of the series was investigated by Tsay, BDS, Keenan and Mc.Leod-Li methods and it was determined that the series were not linear. In the next step, the series were analyzed with nonlinear econometric models in mean and nonlinear econometric models in variance. In the mean nonlinear econometric models, self-excited threshold autoregressive model, smooth transition threshold autoregressive model, logarithmic smooth transition threshold autoregressive model, momentum threshold autoregressive model and Markov switching model were used. Finally, non-linear econometric models of variance such as GARCH, EGARCH, GARCH-M, IGARCH, NAGARCH, GJR-GARCH, TGARCH, AV-GARCH, C-GARCH, APGARCH models were used. As a result, it was seen that unexpected events, arising from internal or external dynamics, triggered regime changes. It has been determined that the structure of investments and the risk perception of investors in the indexes selected from the main stock exchanges of BRICS-T countries differ according to the regimes. It has been determined that volatility is caused by positive and negative news. It has been observed that the effects of negative events on volatility are greater than the effects of positive events on volatility. The relationship between risk and return was found to be statistically insignificant, except for the index selected from India. Findings have been obtained to explain how investors act in the face of changing conditions. Keywords: Market Efficiency, Volatility, Nonlinear Econometric Models in Mean, Nonlinear Econometric Models in Variance