Why Do Banks Fail? - The Explanation from Text Analytics Technique

Abstract

This study investigates the material loss review published by the Federal Deposit Insurance Corporation (FDIC) on the U.S. failed banks from 2008 to 2015. These reports focus on explaining the causes of failure and material loss of each bank. Unlike traditional methods that provide suggestions on financial ratios, our study focuses on phrases extracted from the reports by using text mining technique. Pre-processing steps are used in this study to ‘clean’ the text. Bag of words technique is used for collecting the most frequent words. Topic modelling and document hierarches clustering are used for classifying these reports into groups. Our results suggest that to prevent from being the failure, banks should significantly be aware of: loan, board management, the supervisory process, the concentration of ADC (Acquisition, Development and Construction) and CRE (Commercial real estate). In addition, we find the main reasons that US banks went failure from 2008 to 2015 are covered by two main topics: Loan and Management.

Keywords: text mining, US failed bank, BoW, k-means, topic modelling, hierarches clustering

JEL Classification: G01, G21, G28