台灣總體經濟變數與加權股價指數關係之實證研究

Abstract

本研究探討總體經濟變數與台灣加權指數之關聯性，使用單根檢定、共整合檢定、向量自我迴歸模型、誤差修正模型、衝擊反應函數、預測誤差變異分解以及因果關係檢定等方法來研究時間序列之變數，研究期間為2000年1月至2016年12月之月資料，每筆資料共204個觀察值。實證結果發現總體經濟變數與加權指數皆為I(1)，長期下，具有共整合均衡關係，總體經濟變數對加權指數之衝擊反應函數與預測誤差變異分解影響相當小，並且Granger Causality顯示總體經濟變數與加權指數間亦無顯著之因果關係，表示加權指數有random walk之性質，市場上的訊息早已反應在股票市場上，使用總體經濟變數來解釋加權指數之效果不佳，也可能因為總體經濟變數的資料通常為月資料，對於分秒變動的加權指數不甚敏感。此外，工業生產指數方面，其因果關係顯著落後加權指數，衝擊反應函數與預測誤差變異分解受加權指數與匯率影響劇烈，顯示台灣具有開放經濟體之特質，也證實加權指數為領先指標，用加權指數與匯率來解釋和預測實質產出有相當之解釋力。

This research studies the correlation between Taiwan macroeconomic variables and the stock market index, using unit root test, cointegration test, VAR, VECM, impulse response function, forecast error variance decomposition, and Granger causality test, etc. Period of monthly data is from January 2000 to December 2016. Each variable comprises 204 observations. The empirical results show that Taiwan macroeconomic variables and the stock market index are I (1) with a long-term equilibrium relationship. According to impulse response function and forecast error variance decomposition, response of stock index to macroeconomic variables is very weak. Furthermore, VEC Granger Causality shows that there is no significant causal relationship between stock market index and Taiwan macroeconomic variables, indicating that stock index is like random walk. Stock market reflects all information immediately. Estimating stock market index with macroeconomic variables isn’t effective because monthly data of macroeconomic variables, comparing to stock index that continues changing, is not sensitive. In addition, industrial production index (IPI) is significant Granger caused by stock index. In terms of impulse response function and forecast error variance decomposition, IPI is also significant affected by stock index indicating that Taiwan has the characteristics of an open economy and stock index is a leading indicator. Using stock index to estimate IPI with stock index is effective.