本研究探討在股市中，電子、營建、金融三大類股的交易量之間的關聯性，運用各種時間序列模型，來檢測變數間的關係。 實證結果得到下列發現：(一)單根檢定結果顯示電子、營建、金融三大類股交易量三個變數均為I(1)序列。(二)共整合檢定結果發現各類股交易量間存在著長期均衡的共整合關係，本研究由三種類股彼此間共移(comovement)走勢牽連性強的情況推論，電子、營建、金融三大類股股市的任意組合投資策略都將不具有國內投資組合風險分散的利益。(三)Granger因果檢定得到金融類交易量與電子類交易量具有雙向回饋(feedback)關係，而營建類交易量領先電子類交易量，金融類交易量與營建類交易量則不具因果關係，由此三類股的「領先-落後」關係推論：(1)金融產業與電子產業之間，由「金融海嘯」期間之「三挺政策」可察知兩者之間的相互依存之重要關係，另由兩者產業之間存在互為「借方」與「貸方」密不可分的融資借貸關係來看，正符合金融類與電子類交易量具雙向回饋關係的實證結果；(2)二次政黨輪替後遺贈稅降低為10%的「台商引資」政策，使台商資金回流，大量投入房地產，推升房仲及營建景氣之力道，以至於營建產業取代了電子產業成為領先指，此現象正符合了本研究營建類交易量領先電子類交易量因果檢定之結果。(四)衝擊反應函數分析結果顯示，營建類交易量為一個較獨立的市場，而電子類交易量對自生衝擊的影響，有下降的趨勢，而在金融類交易量方面同時受到電子類與營建類的影響。(五)最後，在本研究變異數分解實證結果得知，三大類股的交易量對自身交易量的變異自我解釋能力都十分強烈，尤以營建類股為最，表示營建類股的外生性很強，不易受他類股的影響，交易市場的獨立性很高。然而，由實證結果發現，電子與金融類股交易量之市場波動產生時，雖短期間受自我解釋能力最強，但在長期互動中，兩類股交易量的自身影響逐漸弱減，而其波動產生的解釋能力，逐漸由營建類股加重對電子與金融類股交易量市場波動之解釋能力。 由本研究結果，充分說明了電子、營建、金融三大類股交易量之間的長短期互動關係，此些結果對於投資人日後進行股市操作，有其重要參考意義。 This study employs various time series models to investigate the relationship among trading volumes of three major stocks of electronics, construction, and finance. There are few empirical findings. First, all three major stocks considered are shown to be I(0) stationary series. Second, there exists a cointegration relationship of long-term equilibrium. The finding of long-term comovement of these three major stocks refers that investors are not able to diversify the risk when portfolio is composed of these three major stocks. Thirdly, lead-lag relationships from Granger causality test show (1) there exists a feedback relationship between stocks of electronics and finance, which reflects the phenomenon when government impose the triple-supporting policy during the sub-prime period, (2) construction stock leads electronic stock, reflecting the recent housing price increases, (3) no causal relation found between financial stock and construction stock. Fourthly, the further findings from the impulse response function illustrate that construction stock is independent of another two stocks, the self-response of electronic stock is in a decreasing shape, and the financial stock is well-responded to the shocks from another two stocks. Finally, the results from the variance decomposition show that construction stock has highest explaining power for the fluctuation occurred in each of the three major stocks. The results from this thesis fully describe the long-run and short-run dynamic relationships among three major stocks of electronics, construction, and finance in Taiwan. We believe that the investors can utilize the findings from our research to set up their future invest strategy.
