عنوان مقاله [English]
نویسندگان [English]چکیده [English]
In this article, we attempt to investigate the factors than effect on price bubbles in Tehran Stock Exchange (TSE) listed companies. First, through runs test, skeweness, kurtosis, and duration dependence test the incidence of bubbles in Tehran Stock Exchange during the years 2004 to 2009 at 95% confidence level were studied. Then all companies that have had severe price volatility have been selected as examples. Totally, a number of 246 companies qualified for the study were selected. And then, by price bubble tests, all companies were divided in two groups, with and without bubbles. To predict bubbles, based on theoretical frameworks, we used indigenous variables such as size of company, P/E ratio, information transparency, stockholder combination, and liquidity rate as independent variables. Then, login binary regression and artificial neural networks (ANN) were used to model and predict bubbles.
Data pertaining to a six-month period prior to formation of bubbles (price acceleration) were employed to fit the model. the results suggested significant relationships between all independent variables and price bubbles. ANN model was identified as a better predictor due to smaller error.