本文結合Jacklin and Bhattachy (1988) 訊息擠兌概念與Diamond and Dybvig (1983) 模型之背景與假設，建構訊息擠兌模型，引發擠兌的因素不再是太陽黑子等隨機變數，而是風險性投資的不利訊息。文中仿效Jacklin and Bhattachrya (1988) 利用特定效用函數，計算出一個主觀機率擠兌臨界值。當低報酬的事後機率值超過此臨界值時，銀行會發生擠兌；反之則不會發生擠兌。該臨界值隨著報酬變異數與相對風險趨避係數的增加而下降，可解釋銀行逾放比率高、擔保品價值波動大、超額貸款以及存款者趨於保守，相對風險趨避係數提高時，都會降低擠兌臨界值，提高擠兌發生機率。本文提出擠兌防護契約，並比較其與一般存款契約的福利大小。

This paper develops an information-based model by incorporating concepts of information and risky investment in Jacklin and Bhattachy (1988) into the Diamond-Dybvig economy. Contrast to bank runs being due to “sunspots” in the Diamond-Dybvig model, bank runs in our model are triggered by depositors’ receiving information indicating a low return on the bank’s risky investments. We derive a threshold rate of bank runs in the sense that when a posterior probability of having a low return on the bank’s investments exceeds this threshold rate, bank runs could be triggered. This threshold rate is shown to be inversely related to the variances of the investment returns and to the attitude of risk aversion. In practice, the threshold rate will be lower and the probability of bank runs will be higher when a bank’s non-performing loans ratio increases, the value of collateral decreases, excess bank lending increases, or depositors are more risk averse. We propose a run-proof deposit contract and compare the magnitudes of depositors’ expected utilities under a run-proof deposit contract and a deposit contract with runs.