摘要
本文嘗試提出一個明確區分內、外在貨幣，且貨幣數量與貨幣政策均為內生的動態隨機全面均衡模型，期望能夠複製出實證上貨幣與產出的動態相關性與統計因果，並為逆向因果假說及支持該假說的實證研究提供具體且可檢驗的理論架構。結果發現，模型大致能複製出台灣實證上貨幣總計數M1與產出的動態相關性，但對於貨幣基數(MB)與產出的動態相關性，本文雖能以央行資訊落後的設定方式來改善模型的解釋能力，卻仍不足以完整詮釋實際資料的特徵。在統計因果方面，本文雖未能以模型中逆向因果傳導機制所產生的模擬值得到實證上“M1 單向Granger cause產出”的結論，但是Granger因果檢定誤判模型中MB與產出間真實的因果關係，的確削弱了傳統上以Granger統計因果來代表真實經濟因果的可信度。此外，藉由實證上貨幣法則轉變的案例來進行模擬比對，本文確認了模型的內生性貨幣政策具有近似中立性的特質。
並列摘要
A dynamic general equilibrium model, in which money and monetary policy are both endogenous, is calibrated and simulated. It is then confronted with the dynamic cross-correlation and statistical causality of money and output in Taiwan time series data. The model economy performs well in matching the empirical dynamic cross-correlation of money aggregate M1 and output. However, it leaves few unexplainable causes of empirical dynamic cross-correlation of monetary base (MB) and output, even with a modification to reflect that the central bank cannot acknowledge the exogenous technology shock immediately. In addition, Granger causality test misjudged the economic causality of MB and output in the model. Therefore, the reliability of such statistical test should be weakened in the aspect of representing real economic causality. Furthermore, through the case in which monetary policy rule is different in two empirical stages, the model proves that endogenous monetary policy is nearly neutral.