本研究主要目的為探討影響台灣地區銀行承做擔保放款契約的決定因素，樣本資料包括38家國內銀行在民國81年底至民國86年6月的四年半期間對國內公開發行公司的923筆放款資料，有效樣本為453筆。經由logit模型的實證分析，主要結果顯示放款額度、到期期間、借款者銷售額、信用風險、借款餘額以及存款餘額等因素對擔保機率產生正向影響；而借款者資產規模愈大以及往來銀行家數愈多者則有較低擔保機率。以整體樣本而言，在借款者的信用評等上發現信用風險較高者對擔保機率具有顯著正向影響，此結果並不支持訊號發射假說所預測高信用品質的借款者會使用較多的擔保契約來發射他們的信用價值至放款市場中。最後，將借款公司分成大小企業的比較結果發現往來關係是決定小企業借款契約擔保機率高低的重要因素。

This paper demonstrates some empirical evidence on the factors that affect the choice of secured versus unsecured bank loan financing. We utilize the logit model to analyze the data on 923 (453 valid) individual loan transactions made by thirty-eight domestic banks to the public companies in Taiwan. The empirical findings indicate that loans of larger size and longer maturity are more likely to be secured. In addition, borrowers with higher sales, credit risk, loan balances and deposit balances tend to use more secured loans; and those with more assets and banks from which the firm borrows are less likely to secure their loans. For the whole sample, bank loans to the firms with lower credit ratings have a significantly higher chance to be secured. This result is not consistent with the signaling hypothesis, which predicts that low-risk borrowers should use more secured loans to signal their credit quality to the loan markets. Finally, a relative comparison between small and large borrowers reveals that banking relationship plays an important role for small businesses in determining whether their loans are secured.