本文研究目的是在動態隨機一般均衡模型當中，加入銀行體系，並以台灣的參數值加以模擬，進一步討論台灣的最適貨幣政策。首先以均衡解進行分析，討論銀行體系如何影響台灣的經濟體，另外，本文探討四種貨幣政策，包含貨幣總數法則，利率法則，其中利率法則包含三種目標：CPI膨脹率目標、國內商品物價膨脹目標及名目匯率目標，模擬結果發現央行使用貨幣總數法則對經濟體的福利大於其他法則，且有助於穩定產出、物價、匯率的波動，另外，提高資本的移動程度以及提高貿易開放程度並不會改變政策福利排序的結果。 / The objective of this paper is to investigate the optimal monetary policy of Taiwan by using a micro-based dynamic stochastic general equilibrium (DSGE) model with the banking sector. The results of steady state show that how the banking sector affects the Taiwanese economy. Furthermore, comparing the performance of alternative monetary policies, the monetary aggregate growth rate rule leads to the highest welfare and lowest volatility of output, inflation rates and exchange rate. It may be the optimal monetary policy of Taiwan. Also, with high degree of capital market friction and highly open market, the rank of these rules is the same.

貨幣政策能否影響實質的經濟活動一直以來都是總體經濟學家討論的主要課題，而欲了解此一課題，便須深入了解貨幣政策影響總體經濟的傳導過程。其中最被人廣為探討的為利率管道與信用管道，而一般而言信用管道可分成兩種基本途徑，分別為銀行借貸管道與資產負債表管道。本文將重點放於銀行借貸管道的探討。 在討論金融因素與經濟體系的互動關係時，既有文獻大多是根據金融加速理論 (financial accelerator theory) 加以探討，該理論認為當實質經濟體系產生干擾時會經由金融體系而使該干擾對經濟的衝擊效果加以放大與加速。因此本文在研究信用管道在我國經濟體系中所扮演角色時，我們採用結構向量自我迴歸模型(Structural Vector Autoregression Model) ，將信用因素作為影響實質經濟體系的干擾來源之一，並分析在1997年亞洲金融危機爆發前後，信用管道與實質經濟之關聯性是否有所不同。另外由於我國過去幾年信用受限制的對象大多是較不具競爭力的傳統產業，或財務出現警訊的企業體，信用管道對經濟體系的影響可能因個人或家計單位，以及民營企業而有所差異。因此本文在討論以本國一般銀行放款所代表的信用變動對經濟體系之影響時，我們將之區分為本國一般銀行對個人與家計單位放款，以及本國一般銀行對民營企業放款兩種不同性質的放款。希望藉由以上的分析，能提供貨幣主管單位一個參考以及喚起讀者對貨幣傳導機制的重視與興趣。

關於貨幣政策傳導機制的探討,包含了傳統的貨幣觀點以及信用觀點,由於傳統貨幣觀點之貨幣政策傳導機制並無法完全解釋經濟變數之變動,因此信用觀點逐漸受到許多學者的重視與討論。信用觀點所探討的貨幣政策傳導機制稱為信用管道,包含了廣義的信用管道(資產負債表管道)與狹義的信用管道(銀行借貸管道)兩種。在本文中,擬從狹義的信用觀點出發,探討不同金融機構是否存在銀行借貸管道。 在本文對信用管道的研究中,我們有別以往以向量自我迴歸模型(VAR模型)、結構化向量自我迴歸模型(SVAR模型)、或將資料一階差分進行一般最小平方法(OLS法)實證研究分析,我們從資料的原始值出發並探討變數之共整合關係(Cointegration),並利用完全修正最小平方法(FMOLS法)對迴歸式進行係數之估計,探討貨幣政策變動下,不同銀行,包括一般銀行、中小企銀、及基層金融機構(農漁會信用部與信用合作社)之放款對於貨幣政策的反應,探討信用管道是否存在於台灣的事實。

無 / This study investigates the monetary effects under the floating exchange rates and imperfect capital mobility by extending the model of Bernanke and Blinder (1988) into a small open economy. It is shown that with credit channel of monetary transmission explicitly considered, the effect of monetary policy on output may be augmented or lessen in our model depending on whether the exchange rate depreciates or appreciates. In addition, the exchange rate puzzle found in the empirical studies can be explained in our theoretical model. The dynamic adjustment patterns of the output and the exchange rate after an increase in money supply are further examined. Under the case of relative high capital mobility, when the real output gradually adjusts toward a higher level, the exchange rate may overshoot, undershoot, or even counter-shoot during the dynamic adjustment process. This provides another one explanation for the volatility of exchange rates under floating rates. Therefore, as financial markets become more internationalized, the conduct of monetary policy turns more complicated in an open economy.

到目前為止，國內外對信用分配的實證研究，大抵不脫以總體資料(aggregate datas)說明在貨幣政策的傳導機制中「信用管道」(credit channel)的存在，但要以此解釋信用分配的現象則略顯不足，我們需要的是更直接的證據：那就是在信用管道背後，銀行用來壓抑地雷公司借款的「篩選工具」(Screening Devices)！ 本文便以分析整理個別公司借款資料的方式進行實證研究，首度將樣本區分為「地雷公司」與「正常公司」，以及「金融風暴前、後」的不同時期，並以「臨界放款利率」作為判斷標準。 本文得到以下三點結論： 1. 在三種樣本期間，地雷公司與正常公司所面對的放款供給曲線，幾乎都有後彎現象（除正常公司在金融風暴前例外），表示信用分配現象的確存在，且銀行以利率為信用篩選工具。 2. 在全部樣本期間與金融風暴後，地雷公司所面對的臨界放款利率均較正常公司為低，表示銀行對地雷公司與正常公司的臨界放款利率的確有異，且銀行成功以「數量」壓抑地雷公司借款。 3. 地雷公司與正常公司在金融風暴後，臨界放款利率均有所降低，且地雷公司臨界放款利率降幅（與全部樣本期間相較）較正常公司高出11個百分點，表示「金融風暴」的確會影響銀行授信態度，且對地雷公司影響較大。 實證結果與研究假說一致，並符合我們的直覺。

In the literature, a considerable theoretical and empirical works have investigated the credit channel of monetary transmission mechanism. This dissertation extends the Bernanke and Blinder (1988) model to an open-economy setting with flexible exchange rate and perfect capital mobility. By means of the framework, we examine the exchange rate dynamics and the adjustment of real output. It turns out that, with a significant credit channel effect, the exchange rate puzzle may occur in the short run and in long run. Moreover, in contrast to Dornbusch (1976), this dissertation shows that, depending upon the strength of the credit channel effect, overshooting, undershooting and counter-shooting impact effect may occur when international capital mobility is perfect. / In the literature, a considerable theoretical and empirical works have investigated the credit channel of monetary transmission mechanism. This dissertation extends the Bernanke and Blinder (1988) model to an open-economy setting with flexible exchange rate and perfect capital mobility. By means of the framework, we examine the exchange rate dynamics and the adjustment of real output. It turns out that, with a significant credit channel effect, the exchange rate puzzle may occur in the short run and in long run. Moreover, in contrast to Dornbusch (1976), this dissertation shows that, depending upon the strength of the credit channel effect, overshooting, undershooting and counter-shooting impact effect may occur when international capital mobility is perfect.

A considerable body of theoretical and empirical literature has evaluated the credit channel of monetary transmission. This paper sets up an open-economy model under floating exchange rates with imperfect international capital mobility based on the Bernanke and Blinder model (1988). Employing our model, we show that a change in money supply has different impacts on the economy in many cases compared to the previous literature. The exchange rate puzzle may occur and when the exchange rate puzzle appears, the Fleming proposition is violated. Besides, by means of a cointegration analysis, we empirically verify the particular case of the exchange rate puzzle with the monthly data from May 1984 to January 2005 in Taiwan. Therefore, our empirical evidences can be matched with our theoretical derivations successfully.