銀行放款行為之順景氣循環現象：機構記憶假說之檢測

銀行業主要功能為提供各種金融中介服務，協助經濟發展與促進景氣繁榮，然而過去十餘年各國學者由不同角度研究發現銀行業可能因各種不同因素造成景氣循環波動擴大，因而提出順循環之現象。本文歸納有關順循環之諸多文獻並引用一個新興發展的理論來分析與探討台灣地區銀行放款的順景氣循環行為，機構記憶假說認為在銀行的放款過程中，放款人員經驗能力的衰退會導致放款標準的寬鬆，進而可能引起銀行放款行為的循環性。以民國76年至96年台灣地區的銀行為樣本，使用panel data方法進行實證分析。研究結果發現機構記憶假說確實能夠解釋台灣地區銀行放款標準的順景氣循環行為，也就是隨著距離上次放款危機的時間越來越遠，存放款利差與擔保放款比率則會逐漸降低；然而，銀行的放款成長率經穩健測試後較無一致性結果。此外，機構記憶假說對小銀行的放款行為有較佳的解釋，且對於仰賴軟資訊類型放款的影響較仰賴硬資訊類型放款的影響來得大。

關鍵字

放款行為；機構記憶假說；順景氣循環

並列摘要

The banking industry provides a variety of intermediary services to assist in economic developments and recoveries. However, many studies of different countries have explored the phenomenon that banks may amplify business cycle volatilities due to various reasons. This is the concept of ”pro-cyclical”. This paper summarizes extensive literature regarding procyclicality, and examines this phenomenon from a new hypothesis that may help explain the procyclicality of Taiwan's bank lending. The institutional memory hypothesis is driven by deterioration in the ability of loan officers over the bank's lending cycle that results in an easing of credit standards, even the procyclicality of bank's lending behavior. We test this hypothesis for the individual Taiwan banks over 1987-2007 and analyze the samples using panel data methods. The results show that the institutional memory hypothesis is able to explain the procyclicality of Taiwan's bank lending behavior. It means that loan spread will reduce as well as the proportion of secured loans as time passes since a bank's last bust. However, the robust test appears to provide no consistent evidence for loan growth. The findings also indicate that the institutional memory hypothesis explains better the lending behavior at small banks than large ones. Finally, as expected, the institutional memory hypothesis is more supported by soft-information-intensive loans than hard-information-intensive loans.