Research on correlation between listed stock prices and ADR prices with Haar Wavelet application

碩士 === 銘傳大學 === 財務金融學系碩士在職專班 === 94 === Abstract By adopting Haar Wavelet Model, this research transforms underlying stock and ADR prices into several different scales and investigates their stationarity and long-term relationship by using Johansen Cointegration Test. Besides, based on Granger Caus...

Full description

Bibliographic Details
Main Authors: May-Lin Wu, 吳美霖
Other Authors: Chong-Jen Yang
Format: Others
Language:zh-TW
Published: 2006
Online Access:http://ndltd.ncl.edu.tw/handle/ry64hp
Description
Summary:碩士 === 銘傳大學 === 財務金融學系碩士在職專班 === 94 === Abstract By adopting Haar Wavelet Model, this research transforms underlying stock and ADR prices into several different scales and investigates their stationarity and long-term relationship by using Johansen Cointegration Test. Besides, based on Granger Causality Test of different scales, this study leads result of up-and-down relationship between underlying stocks price and ADR price to certify the hypothesis of “ Dominant and Satellite Market” from Garbade and Silber. According to empirical results, our finding can be summarized as: 1. Unit Root Test: In the test of 7 stock prices and ADR prices, it would be permitted none of unit root existence as long as original series transform though first difference. On the other hand, most of wavelet transformed time series are permitted no unit root without any difference, while the last scale one is permitted no unit root after taking second difference. 2. Co-integration test: In the test of underlying of 7 stock prices and ADR prices, we found original series of factors consist of different long-term relationship. In additions, after Haar Wavelet transforming, different scales series result in various and more detailed co-integration relationship beyond original series providing. 3. Granger Causality Test: In the test of underlying 7 stock prices and ADR prices, besides up-and-down sense from original series, we observed clearer trend among factors in different scales series.