896期 5月17日 :Bank Runs and Regulatory Communication: An Experimental Analysis(Authors: Surajeet Chakravarty, Miguel A. Fonseca, and Todd R. Kaplan)(Todd R. Kaplan, 教授,University of Exeter and University of Haifa)
上海财经大学经济学院
科学研究
当前位置:首页 >> 科学研究 >> 学术讲座
896期 5月17日 :Bank Runs and Regulatory Communication: An Experimental Analysis(Authors: Surajeet Chakravarty, Miguel A. Fonseca, and Todd R. Kaplan)(Todd R. Kaplan, 教授,University of Exeter and University of Haifa)

【主讲】Todd R. Kaplan (教授,University of Exeter and University of Haifa)

【主题】Bank Runs and Regulatory Communication: An Experimental Analysis(Authors: Surajeet Chakravarty, Miguel A. Fonseca, and Todd R. Kaplan)

【时间】2018年5月17日 (周四) 15:30-17:00

【地点】上海财经大学经济学院楼701室

【语言】英文

【摘要】The stability of financial systems has been at the forefront of policy-makers after the world witnessed the collapse of major financial institutions in the 2007-2008 financial crisis. A better understanding of the causes of bank failures is essential to avoid the significant welfare losses witnessed. We experimentally analyze the role that revelation of financial information plays in the likelihood of having a bank run. We base our experiments on the Diamond and Dybvig (1983) model. We have two depositors that have to decide to withdraw today or tomorrow. If both withdraw today, they both receive 200. If they both withdraw tomorrow, they both receive R, which is randomly selected a set of numbers that range from 80 to 720. If one withdraws today and the other tomorrow, the one withdrawing today receives 400 and the one withdrawing tomorrow receives 0. A regulator knows the value of R. In our TRUTH treatment, the regulator must send the precise value of R to the depositors before they decide. In our ANYT treatment, the regulator can send any range of R as long as it contains the true value of R. In ANY, the regulator can send any range of R. We find that Depositors' early withdrawal decisions as a function of the lower bound of the message about R are significantly different in each of the three treatments. There is a significant drop in the likelihood of withdrawing early in ANY and TRUTH when the lower bound of the message equals 400. In addition, we find that there are more inefficient runs (for large R) and fewer efficient runs (for small R) in ANY than in either TRUTH or ANYT. In ANY and ANYT, there were heterogeneous strategies used by the regulators. Some were strategically vague and hid worse states with good states. In a second set of experiments, we determine which strategy is the best for the regulator.

上一条 | 下一条
上海财经大学经济学院