台灣上市加權股價指數與雙率之關聯性分析

Study on the Relationship Among Taiwan Stock Index,Discount Rate, and Exchange Rate

徐衍瑜(HSU, YEN-YU)

指導教授 ： 陳達新

國立臺北大學/商學院/國際財務金融碩士在職專班/碩士(2011年)

摘要

當前金融市場情勢大幅演變，導致金融商品市場價格波動劇烈，股票市場經過幾翻波折後，投資者開始注意到股票市場崩解價跌的風險，如何建立分散風險的投資組合，成為當前投資市場重要課題。
本研究在探討台灣上市加權股價指數與新台幣匯率、公債、短期票券利率、新台幣利率交換之關聯性，運用單根檢定、Johansen共整合檢定、向量自我迴歸模型、向量誤差修正模型及Granger 因果關係進行實證分析。變數資料選取範圍自 2000年1月至2010年12月之日資料，採用台灣發行量加權股價指數、新台幣匯率、台灣十年期公債殖利率、台灣五年期公債殖利率、台灣五年期利率交換與台灣九十天期商業本票利率等變數進行實證分析，實證結果如下：
一、各變數在經由ADF 單根檢定下，均具有單根的性質，顯示均為非定態，唯在經由一階差分後各變數皆在 1% 顯著水準下，拒絕時間序列具有單根存在的虛無假設，表示各序列皆為定態而且具有相同整合級次 I(1) 的型態。
二、共整合檢定在 1 %的顯著水準之下，各變數間存在一組長期穩定均衡關係的共整合向量，因此顯示各變數間在長期下皆存在穩定的均衡關係。
三、台灣發行量加權股價指數和台灣十年期公債殖利率間、台灣發行量加權股價指數和台灣五年期公債殖利率間、台灣發行量加權股價指數和台灣五年期利率交換間，受到金融市場間之利差水準、匯率的走勢進而影響台股的走勢，造成相互牽動彼此的走勢，因此形成互有領先之雙向回饋關係。
四、當各變數發生自發性變動時，皆對本身在短期內的衝擊影響最大。對其它的變數的衝擊以台灣五年期公債殖利率的正向反應最大，衝擊反應屬於長期型態，長期累積則具正向效果。因此在短期台股指數時，除以參考本身之落後期之外，仍可觀察股市波動對公債的影響。
五、預測誤差變異數分解：各變數之變異數由自身解釋的能力均相當的高，台灣發行量加權股價指數之自發性高，較不易受外生變數的影響，不過在長期而言，股市、貨幣市場與匯市則展現緩升的影響能力。

關鍵字

單根檢定 ； 共整合 ； 向量自我迴歸 ； 因果關係 ； 預測誤差變異數分解

並列摘要

When situations of current financial markets have developed significantly, leading to drastic fluctuations in the market price of financial products, investors have begun to notice price risks before the collapse of the stock market, with times of upheavals from the stock market. Therefore, how the risks of investment portfolios can be diversified has become an important issue for current investment markets.
This s tudy aims to investigate the interrelationship between the weighted stock market index of Taiwan to the exchange rate of the New Taiwan dollar, government bonds, short-term bill interest rates, and interest rate exchange of the New Taiwan Dollar, making use of the unit root test, Johansen co-integration verification, vector auto-regression model, vector error correction model, and Granger causality to conduct substantive analysis. Variable information is selected from data between January to December of 2000, using such variables as the Taiwan TAIEX Index, New Taiwan Dollar exchange rate, yield of 10-year Taiwan government bonds, yield of 5-year Taiwan government bonds, Taiwan 5-year interest rate exchange, and interest rate of Taiwan 90-day commercial paper to conduct empirical analysis, with results obtained as follows:
1.Under ADF unit root verification, each variable is found to have unit root, showing that they are non-stationary. However, each of the variables is found under 1% prominent level as conducted after first difference, denying virtual assumption that time series is found with existence of unit root, and it indicates that each of the series is both stationary and found with similar mode of integration grade I(1).
2.For test of co-integration under the 1% prominent level, each variable is found with co-integration vector with a set of long-term stable and stable equilibrium relationship, and it is why it shows that a stable and equilibrium relationship is found existing among each variable for a long period of time.
3.It is found that the relationship between Taiwan TAIEX Index to the yield of 10-year Taiwan government bonds, Taiwan TAIEX Index to yield of 5-year Taiwan government bonds, and Taiwan TAIEX Index to the 5-year interest rate exchange of Taiwan is affected by interest rate levels and fluctuations in exchange rates among financial markets, thus affecting fluctuations of stocks of Taiwan. As a result, they have, mutually, affected the fluctuations of each other, thus forming a two-way feedback relationship of leading in their own domain.
4.When each variable is found with spontaneous change, it will exert the greatest impact on them in a short time-span. As for impact on other variables, it exerts the greatest positive response upon the yield of Taiwan government 5-year bonds, and the impact is considered to be perennial and it will perennially accumulate a positive effect. Therefore, in the short-term the TAIEX Index can, aside from reckoning its backwards phase, also be used to observe the impact of stock fluctuations on government bonds.
5.Forecast error variance decomposition: the capability of the variance of each variable to explain itself is considered rather high, and the spontaneity of the TAIEX Index is very high, so that it is less likely to be affected by external variables. However, the stock market, currency market, and exchange market will, over the long term, demonstrate the capability of influence for slow rise.

並列關鍵字

unit root test ； co-integration ； vector auto-regression ； causality ； forecast error variance decomposition

參考文獻

24.簡雨柔，2007，「考慮流動性下之利率期間結構預測」，淡江大學財務金融學系，碩士論文。

12.Jarrow, R., David Lando and Stuart Turnbull, 1997, “A Markov Model of the Term Structure of Credit Spreads”, Review of Financial Studies 10(2), Summer,77-82

3.王運強，2005，「台灣公債市場與台幣利率交換交易市場動態關聯性之研究」，國立中央大學財務金融學系碩士在職專班，碩士論文。

4.謝文馨 ，2008，「總體經濟變數與股價指數之關聯性研究─以台灣為例」，國立成功大學企業管理學系，碩士論文。

8.李亞純，2007，「公司債信用價差分析-台灣與美國債券市場之實證探討」，淡江大學財務金融學系，碩士論文。