Stylized Facts of Financial Markets and Modeling Volatility (The Case of Tehran Stock Exchange)

Abstract

The financial markets stylized facts, volatility and its relationship with returns tested empirically in Tehran Stock Exchange (TSE). ARMA- ARCH type models including two symmetric conditional hetroscedastic models; ARMA (1, 1) - ARCH (1) and ARMA (1, 1) - GARCH (1, 1) and two asymmetric conditional hetroscedastic models; ARMA (1, 1) - GJR-GARCH (1, 1) and ARMA (1, 1) - EGARCH (1, 1) are used to determine the process of stock returns and volatility, and by applying ARCH(M) class models and out-of-sample methodology the relationship between stock return and volatility is examined. The key findings are as follows, In contrast to emerging markets which shows high volatility, TSE shows less volatility and it is due to unique features of TSE, Volatility Range, Basic Volume and Knot Event. Heavy tails, volatility clustering and mean reverting behavior are observed in TSE, however, the leverage effect is not observed. Estimates of model parameters suggest that the symmetric models explain the most process envisaged in the time series data, while the asymmetric models fail. TSE fails to confirm the weak form of efficient market hypothesis as returns have serial correlation, thus it supports the idea that controlling mechanism of Volatility Range and Basic Volume should be removed in TSE to attract foreign investment and finally we find no evidence of relationship between return and volatility.