Abstract:
The research problem investigated in this paper is modeling volatility and analyzing risk and return’s relationship in Tehran Stock Exchange using GARCH-family models including GARCH(1,1), GARCH(2,2), EGARCH(1,1), PGARCH(1,1), TGARCH(1,1), GARCH(1,1)-M and CGARCH(1,1). Using the daily returns of Tehran Stock Exchange companies, we focused on two portfolios of all the companies during a 10-year-period and those liquid ones during a 5-year-period. In order to meet the distributional characteristics of the financial time series, we have used Normal, t-Student and GED distributional assumptions. The results of this survey and applied research show that first, the conditional volatility models best succeed in modeling characteristics of financial data including volatility clustering, long memory and leverage effects. Second, for both portfolios, increased risk will lead to a rise in the returns.
Machine summary:
Master of Science Student in Financial Management, University of Tehran, Iran (Date of article received: 41/01/78, Date of approval: 61/06/88) Abstract The issue under investigation in this research is modeling volatility in the Tehran Stock Exchange and analyzing the relationship between risk and return using GARCH family models.
The results of this research, which is of a survey and applied type, indicate that firstly, conditional variance heterogeneity models can well model the characteristics of financial data such as clustered volatility, long-term memory, and leverage effects.
The results of model estimation regarding the market portfolio data are as follows: In general, the positive and significant values of ?, ?, ?, and ?
in the GARCH(2,2)-N model, which is also statistically significant, indicates that the conditional variance is negatively correlated with the values of shocks from two periods ago.
High values of ?, which are close to one in the models of GARCH(1,1)-G, PGARCH(1,1)-T, PGARCH(1,1)-G, EGARCH(1,1)-T, and EGARCH(1,1)-G, indicate the long-term impact of shocks on conditional variance, a phenomenon that points to the existence of long-term memory in market returns.
The results of estimating the models for the series of returns of the portfolio of fifty companies are as follows: Positive and significant values of ?
The results of estimating the GARCH(1,1)-M model show that ?s are significant in all distributions under investigation, which proves the existence of a positive correlation between risk and return.
The results of estimating the GARCH(1,1)-M model show that ?s are significant in all distributions under investigation, which proves the existence of a positive correlation between risk and return.
(
, 
, 
)
