The Study on Herding Behaviors among Banks for Industrial Firms$ Loans in Taiwan

This plan extends definition of institutional herding proposed by Sias (2004) to examine whether the dynamic herding behaviors exist among the bank industry in Taiwan with regard to the industrial firms’ loan, what the cause of their herding in granting a loan is and whether their herding of providing a loan significantly increases in the bullish period comparative to in the bearish period and this phenomenon mainly results from pursuit of the same type of bank loan. By testing the cross-section of the bank industry towards loan demand of the industrial firms in any two adjacent semiannual periods and decomposing the cross-sectional dependence into their own cascades and others cascades (i.e. capital loan herding), I examine whether the cascades of the bank industry in Taiwan towards the industrial firms’ loans mainly results from their herding behaviors. Concurrently, by using the cross-sectional correlation that the bank increases the weight rate of their loan portfolio for adjacent inter-semiannual periods, this plan plans to analyze whether loan herding among the banks is the result of the habit loan. Next, I would like to examine whether the loan made by a bank to the industrial firm are more likely to herd in the industry with larger or smaller average capitalization; thus, “investigative herding” or “information cascades” are the main reasons of loan herding among the bank industry in Taiwan towards the industrial firms. Moreover, this plan examines whether there is a more significant herding phenomenon of the enterprise loan in the bullish period comparative to in the bearish period in the stock market among the bank industry in Taiwan. In addition, this plan examines whether the enterprise capital loan made by all types of bank in Taiwan are more likely to follow similar types of bank rather than different types of bank on the one hand, the result of which may be consistent with the view proposed by Sias (2004) that institutional herding is driven by reputation herding. On the other hand, the author also examines whether the small-size banks will also significantly follow the large-size banks in the enterprise capital loan and whether public banks will also significantly lead private banks, the result of which may be consistent with the argument of Del Guercio (1996) and Bennett, Sias and Starks (2003) that institutional herding is driven by characteristic herding.[[abstract]]本計畫延伸Sias (2004)對法人群聚的定義，以檢測是否臺灣銀行業之間對產業廠商的放款存在動態的群聚行為，什麼是其放款時群聚的原因，是否其放款的群聚現象於多頭相對空頭期間明顯增加且是否該現象主要源於追隨相同類型銀行放款。藉由檢測任兩個鄰近半年銀行業對產業廠商放款需求的橫斷面相依並分解橫斷面相依為自身串流及他人串流(即放款群聚)，我檢視是否臺灣銀行業對產業廠商放款的串流主要源自於其群聚行為。同時，藉由使用銀行增加其跨鄰近半年放款組合權重比率的橫斷面相關，本計畫擬分析是否銀行之間放款的群聚是習慣性放款的結果。接續，我想要檢測是否銀行對產業廠商放款較可能群聚於平均資本額較大或較小的產業，因而”調查群聚”或”資訊串流”是台灣銀行業之間對產業廠商放款群聚的主要原因。再者，本計畫檢測是否台灣銀行業之間對企業放款於股市多頭期間相較空頭期間有更明顯的群聚現象。此外，本計畫一方面檢測是否台灣各類型銀行對企業資金貸放更可能追隨相同類型銀行而非不同類型銀行，此結果可能與Sias (2004)提及”法人群聚由聲譽群聚所驅動”的觀點有一致性。另一方面，本計畫亦檢測在企業資金貸放上是否台灣小規模銀行也會明顯追隨大規模銀行,且是否公營銀行也會明顯領先民營銀行，此結果也可能與 Del Guercio (1996) 和 Bennett, Sias and Starks (2003) 提及”法人群聚由特徵群聚所驅動”的論點相一致