總體經濟景氣循環變數與股市報酬之相關性-以台灣為例

Abstract

本研究係探討總體經濟之景氣循環變數與股價指數的相關性，樣本期間為民國81年1月至年95年3月，選取台股大盤指數與景氣同時指標(製造業生產指數、製造業銷售值、製造業員工平均每月薪資、票據交換金額、國內貨運量)，欲建立VAR模型，並進行景氣循環及股市之因果檢定。 本研究發現，景氣循環同時指標和台灣股市間具有共整合，短期可能會受外生變數的衝擊產生短暫偏離，但長期會回復均衡狀態，在考慮長期資訊下，需使用誤差修正模型VECM來進行因果分析，結果顯示，股市價格和景氣循環同時指標皆有顯著因果關係，股市價格顯著領先景氣循環變數。此外，股市變動會領先景氣循環變數約5∼6個月，且長期皆為正向影響。 本文第一章將先說明研究動機和架構，並於第二章進行相關文獻的討論，第三章敘述採用的研究方法，第四章則是實證資料分析，最後為結論和建議。

This study discusses the relation between the macroeconomic business cycle variables and the stock prices. We choose five coincident indicators of business cycles and the sample period is from January 1981 to March 2006. In addition to vector autoregression (VAR) model, we also test the causality relations between the businesses cycle and the stock market by the Granger causality test. One of our results is that the cointegration relation existed between theses variables. In the short run, the exogenous factors may impact the equilibrium hence the temporary bias occurs, but it will return to the equilibrium in the long run. Considering the long-run situation, we have to adopt error correction model (VECM) to analyze the causality relations. It shows that stock prices have significant Granger causality relations with all five coincident indicators, manufacturing production index, manufacturing sales, wages in manufacturing, amount of banks’ clearing checks & bills, and quantum of domestic traffic. Moreover, stock prices lead the business cycles coincident indicators about half a year in Taiwan.