自2003年臺灣發行第一件資產證券化商品以來，整體證券化市場發展快速，在可預見的未來，臺灣資產證券化將迅速普及。不過，就一般而言，間接金融與直接金融間似有替代關係，亦即資產證券化可能產生金融逆中介現象。因此，本文的研究目的即在於探討資產證券化對臺灣銀行業放款之影響為何。 　　本文以2001年第4季至2004年第4季臺灣銀行業將其資產予以證券化的不平衡追蹤資料，搭配固定效果模型的估計，實證研究發現，從臺灣短暫的資產證券化經驗來看，資產證券化的增加確實會造成銀行放款量的相對減少（放款餘額占資產之比率下降），即臺灣實施資產證券化後，產生金融逆中介現象。長期而言，直接金融（含資產證券化）的增加將造成銀行放款成長率趨緩，以及銀行放款餘額占資產之比率下降，但此並不意味資產證券化的普及，將造成銀行總放款量減少，而降低銀行體系金融中介的功能。 　　另外，透過固定效果模型，發現其他因素對銀行放款之影響如下：（1）銀行逾期放款比率雖與放款呈負向關係，但並不顯著。意即銀行在逾期放款比率增減時，並不會特別調整放款餘額占資產之比率。（2）銀行淨值與放款之關係亦不顯著。表示銀行淨值增加時，放款可能只會與其他資產同步增加，而不會因此特別擴大其占資產之比率。（3）房價指數的係數為顯著正值。顯示擔保品價值愈高時，銀行愈願意辦理放款。（4）其他直接金融存量（扣除資產證券化）的係數為顯著負值。顯示直接金融與間接金融存在明顯替代關係。（5）國內生產毛額與銀行放款之關係不顯著。表示國內生產毛額增加時，銀行放款可能只會與銀行其他資產同步增加，而不會因此特別擴大其占資產之比率。 　　在控制其他解釋變數後，計算各樣本銀行的特質效果（即放款餘額占資產之比率），發現除了放款餘額占資產之比率較高者，逐漸降低該比率外，其餘銀行並無特定趨勢。因此，似可推論出銀行放款餘額占資產比率的高低，與銀行本身之經營策略有關，致使該比率在不同銀行間有不同水準。 / The entire securitization market has been evolving rapidly since the first asset securitization product was issued in Taiwan in 2003. In the foreseeable future, asset securitization in Taiwan will become prevalent. However, there seems to be a substitution relationship between direct finance and indirect finance; that is, financial disintermediation may emerge as a result of asset securitization. Therefore, this paper aims to discuss the effects of asset securitization on bank loans in Taiwan. Based on the unbalanced panel data of the asset securitization of banks in Taiwan from Q4 2001 to Q4 2004 along with the estimates from the fixed-effects model, it is found in this study that, judging from Taiwan’s brief experience in asset securitization, an increase in asset securitization does indeed bring about a relative decline in the amount of loans (a decreased ratio of loan balance in assets). In other words, financial disintermediation has arisen with asset securitization in Taiwan. From a long-term perspective, increment of direct finance (including asset securitization) will lead to retarded growth in bank loans as well as a lower ratio of loan balance in assets. This, however, may not necessarily imply that the popularization of asset securitization would result in a decrease in the amount of bank loans or weaken the financial intermediation function of the banking system. In addition, effects of other factors on bank loans found via the fixed-effect model are as follows: (1) Non-performing loan (NPL) ratio and bank loans have an inverse relationship, albeit insignificant. This means that banks do not usually adjust the ratio of loan balance in assets in accordance with their NPL ratios. (2) The relationship between the net worth of banks and the amount of loans is insignificant as well. This indicates that the amount of loans would only rise with other assets as the net worth of banks increases; the ratio of loan balance in assets will not be specifically elevated. (3) The coefficient of Housing Price Index is significantly positive, indicating that the higher value a collateral has, the more a bank is willing to release a loan. (4) The coefficient of other stock of the direct finance (excluding asset securitization) is significantly negative, which reveals an obvious substitution relationship between direct finance and indirect finance. (5) The relationship between gross domestic product (GDP) and bank loans is insignificant, which indicates that the amount of bank loans would only rise with other bank assets as GDP increases; the ratio of loan balance in assets will not be elevated accordingly. As other explanatory variables are under control, the results gained from computing the specific effects (the ratio of loan balance in assets) of each sample (bank) show that only those with a higher ratio of loan balance in assets are found gradually reducing the radio. Such a trend is not found in others. Consequently, it can be inferred that the ratio of loan balance in assets depends on the business strategy of the bank itself, which results in different levels of the ratios among different banks.