[[alternative]]A Study on the Relationships Between Stockindex and Interest Rate by using Vector Autoregression Model

[[abstract]] 美國經濟是世界經濟的櫥窗，其經濟政策的改變影響所及，往往不只是美國經濟的表現，與其依存度高的國家也都受連帶的影響，台灣當然也不例外。又一個國家股價指數的高低常可反映該國經濟情況的好壞，故本文乃透過ARIMA、GRANGER因果關係檢定、VAR模式、共整合關係檢定、變異數分解以及衝擊反應，分析台灣之股價、重貼現率、隔夜拆款利率與美國之股價、重貼現率等，五個變量之間的關聯性，以了解美國利率政策情況對台灣股價與利率的互動關係，其研究結果如下： 1. 經由共整合檢定得知，上述五個變量不具有長期均衡關係，由衝擊反應分析可知，當此五個變量分別發生自發性干擾時，其衝擊反應效果由第14期至第20期之後才會消失。 2. 由GRANGER因果關係檢定及變異數分解之結果顯示，台灣重貼現率受美國重貼現率之影響較受台灣股價影響大；雖然台灣之重貼現率與隔夜拆款利率為回饋關係，但重貼現率受隔夜拆款利率影響較大；此外美國重貼現率影響台灣重貼現率之程度較影響台灣隔夜拆款利率大。 3. 由VAR及ARIMA模式比較，台灣、美國股價、台灣重貼現率與隔夜拆款利率之ARIMA模式的配適度較VAR模式佳，表示此四個變量由過去本身歷史資訊即可反應本身的走勢；至於美國重貼現率之VAR模式較ARIMA模式來得佳，即美國重貼現率的水準可由自身及股價指數之歷史資料來反應。[[abstract]] America’s economic is the module of world. The changing of financial policy in the U.S. not only affects America’s mo netary performance but also the countries are interdependent with the U.S. very much. And Taiwan is one of them. Also, the stock indicts usually means one country’s economic status. In this research, there are 6 methods (ARIMA, Granger causality test, VAR Model, cointegration test, variance decomposition and impulse response) used to analyze the connection among 5 variables which are Taiwan’s stockindex, discount rate, interbank offered rate, America’s stockindex and discount rate. In addition to, we did research the relationship between the stockindex and interested rate in Taiwan when America adjust the interested rate police. Finally, we got the critical result herein. At first, by co integration test, these 5 variables do not have long-term equilibrium. Differently, the impulse response will disappear during 14th to 20th when the items have spontaneous interference individually by impulse response theory. Second, the result shows the discount rate in Taiwan is affected by America’s discount rate more than Taiwan’s stockindex by using Granger causality test. The discount rate has feed back relationship with the interbank offered rate. However, the truth is that the discount rate is mainly affected by the interbank offered rate more. Except, the influence of interbank offered rate is much more than discount rate in Taiwan by America’s discount rate. The Third, we compare the stockindex, discount rate, interbank offered rate in Taiwan and stockindex in the U.S. we find the fit of ARIMA is better than VAR model. It means the four variables can respond their self trend by their own earlier period information. In the other side, VAR model is better than ARIMA model for America’s discount rate. That tells us the standard of America’s discount rate can be responded by the historical data of America’s discount itself and stockindex