Summary: | 碩士 === 國立臺北大學 === 國際財務金融碩士在職專班 === 105 === This paper adopts high frequency intraday data to investigate the impacts of the variation of order book on stock price. In particular, the study attempts to find out a real time trading model which could enable investors, based on the best 5-set order books microstructure, to make prompt analysis and judgement on daily investment decisions. The intent is to make investment easier by glancing the disclosure book in Taiwan stock market.
The objectives of this paper are to investigates the impacts of various bid-ask volume differentials on stock price, to examine the discrepant effects during different trading hours on a trading day, and to study the asymmetric impacts between the bull and bear trading strategies. Three underlying stocks were chosen, i.e., Taiwan Top 50 ETF (0050 TT), CSI 300 ETF (0061 TT), and LARGAN (3008 TT). The second-by-second intra-day data, ranging from 2015/07/01 to 2016/11/16, a total of 340 trading days and 960,000, 630,000 and 540,000 observations, were selected respectively. The logistic regression model was utilized for estimation and statistical tests.
The result shows that the average model hit ratios for the three stocks (#0050, #0061 and #3008) were 94.66%, 91.21% and 94.25%, respectively. In terms of trading hours, the best hit ratios for long is located at 12:00~13:00, and the worst was between 9:00~9:30. Besides, the highest and lowest hit ratios for short were different in three underlying stocks. The positive returns in #0050 was located at 9:30~13:30, in #0061 was located at 9:00~9:59 and 10:30~13:30, and in #3008 was located at 9:00~9:59 and 12:30~13:30.
|