| Summary: | 碩士 === 國立雲林科技大學 === 財務金融系碩士班 === 94 === The purposes of the present paper are four folds: (1) Examine price manipulation on the settlement dates of TAIFEX TAIEX futures (TAIFEX TF) and SGX MSCI Taiwan Index futures (SGX TF). (2) Inquire the predictability on the binary directions of either pulling up index or squeezing down index on the settlement dates via an industry based model. (3) Survey the effects of the settlement institutional changes on the alienation of the price manipulation phenomenon. The empirical data used in this study spans 1998.7.22-2005.10.19, covering the available eighty six settlement dates. The main evidences on this study are: (1) there do exist the expiration date effects for TAIFE FX and SGX TF, in particular, MGX TF showing strong significance in the empirical test. (2) The industry based five factors model renders over 70% prediction accuracy via the Logistic regression and over 80% for a feedforward artificial neural network. Moreover, the predictor variable of the previous price change of US Dow Jones Industry Average Index exhibits a pompous explaining power to the binary direction of price manipulation. (3) Changes in settlement mechanisms do improve the price volatility in the settlement dates for both TAIFE FX and SGX TF. Specifically, after adopting the fifteen minutes average prices rule, TAIFEX TX has significant reduced its volatility since then. Meanwhile, TAIEX taking pre-close five minutes call price also has been improving the price volatility on MSCI TF settlement dates. Both mechanism dummy variables have significantly improved the price manipulation on futures settlement dates.And(4)Simulation of trading profit shows that the market model gains significantly higher profit than a naïve trading strategy does.
|
|---|
