本文應用一種結合頻率範疇的頻譜帶濾波分析及時間範疇的向量自我迴歸模型的 研究方法，探討臺灣股市報酬與貨幣成長及物價膨脹等總體經濟變數間跨頻譜帶之動態因果 關連；另外並透過交叉頻譜分析確認其間之領先及落後情形。研究結果發現（１）貨幣成長 的短期（１至２個月）及中長期（２０至２８個月）波動會對股市報酬有正向的影響，然股 市報酬的波動卻不會顯著影響貨幣成長的變動；（２）股市報酬與物價膨脹間的困果關係並 不顯著；（３）就領先及落後的情形來看，貨幣成長領先股市報酬的變動為半個月至一個月 ，而股票報酬則領先物價膨脹變動一個月。

　　　　　This study applies a quantitative approach which integrates band spectrum analysis and vector autoregression (VAR) to analyze the dynamic causal relationship between stock returns and macroeconomic variables. The lead and lag relationships between the variables are also explored by using cross-spectral analysis. The results indicate that 1) short-run (1 to 2 months) and medium-long-run (20 to 28 months) fluctuations in money growth will have positive effects on stock returns, while stock return fluctuations have only minor effects on money growth; 2) there are no significant relations found between stock returns and inflation; 3) money growth moves about one half to one month ahead of stock returns, while stock returns lead inflation by about one month.