社會科學院: 經濟學研究所指導教授: 毛慶生周士凱Chou, Shih-KaiShih-KaiChou2017-03-032018-06-282017-03-032018-06-282016http://ntur.lib.ntu.edu.tw//handle/246246/275265本文使用2005:W1至2015：W43之週資料，共564筆週資料觀察值，並運用多種時間序列計量方法，以探究美元指數期貨投機者淨部位是否能預測美元指數期貨走勢。 利用向量自我迴歸模型探討美元指數期貨之投機者淨部位與美元指數期貨走勢之間關聯性。直觀上，投機者淨部位是經過專業投資人士蒐集並研究而下單產生的交易留倉結果，可能可以領先市場探知美元指數期貨未來走勢，本文研究結果顯示，美元指數期貨投機者淨部位無法預測美元指數期貨市場未來走勢，反而是美元指數期貨走勢對美元指數投機者淨部位具有解釋力，美元指數投機者大多仍是採取順勢交易的方式（Trend Following）。This paper uses weekly data from 2005:W1 to 2015:W43, covering 564 entries of observation data, to explore whether net positions of Non-commercial traders in dollar index futures market can predict the return of dollar index futures market. The correlation between two variables is determined with the vector autoregression model (VAR). Intuitively, net positions of Non-commercial traders in dollar index futures market is the result of professional investor studies and could potentially predict the return of dollar index futures market. However, there is substantial evidence that traders respond to price change and Non-commercial traders display a tendency for trend following. Second, there is practically no evidence that traders’ positions can forecast or lead market returns.527013 bytesapplication/pdf論文公開時間: 2021/3/8論文使用權限: 同意有償授權(權利金給回饋學校)單根檢定向量自我迴歸模型Granger因果關係檢定衝擊反應分析預測誤差變異數分解unit root testvector autoregression modelGranger causality testimpulse response analysisforecast error variance decompositionthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/275265/1/ntu-105-P01323005-1.pdf