Financial Information Asymmetry: Using Deep Learning Algorithms to Predict Financial Distress

Chyan-Long Jan

doi:10.3390/sym13030443

Abstract

Because of the financial information asymmetry, the stakeholders usually do not know a company’s real financial condition until financial distress occurs. Financial distress not only influences a company’s operational sustainability and damages the rights and interests of its stakeholders, it may also harm the national economy and society; hence, it is very important to build high-accuracy financial distress prediction models. The purpose of this study is to build high-accuracy and effective financial distress prediction models by two representative deep learning algorithms: Deep neural networks (DNN) and convolutional neural networks (CNN). In addition, important variables are selected by the chi-squared automatic interaction detector (CHAID). In this study, the data of Taiwan’s listed and OTC sample companies are taken from the Taiwan Economic Journal (TEJ) database during the period from 2000 to 2019, including 86 companies in financial distress and 258 not in financial distress, for a total of 344 companies. According to the empirical results, with the important variables selected by CHAID and modeling by CNN, the CHAID-CNN model has the highest financial distress prediction accuracy rate of 94.23%, and the lowest type I error rate and type II error rate, which are 0.96% and 4.81%, respectively.

Highlights

In the latter half of 2008, the subprime mortgage crisis in the United States continued and spread worldwide, triggering a global financial crisis and affecting the global economy
Comparison of the performance of the four financial distress prediction models built in this study as shown in Table 4: The chi-squared automatic interaction detector (CHAID)-convolutional neural networks (CNN) model has the highest accuracy, followed by the CNN model, the CHAID-deep neural networks (DNN) model, and the DNN model
According to the empirical results by the accuracy of test dataset: (1) The four financial distress prediction models built in this study all have very high accuracy, which is above 89%; (2) CNN has a higher prediction accuracy than DNN; (3) important variables are first selected by CHAID, and input into DNN and CNN to build the models, which are more accurate than the 23 variables directly input into DNN and CNN to build models without selection; (4) for the four financial distress prediction models, the type I error rates are all below 6%, and the type II error rates are below 6%, indicating that the four models are all effective

Summary

Introduction

In the latter half of 2008, the subprime mortgage crisis in the United States continued and spread worldwide, triggering a global financial crisis and affecting the global economy. Impacted by the subprime mortgage of the United States, in 2008, the Indy Mac Bank was in financial distress, and was taken over by the Federal Deposit Insurance Corporation of the U.S government, Lehman Brothers declared bankruptcy, and the American International Group (AIG) applied to the Federal Reserve Board (FRD) for emergency financing of USD 85 billion. The global financial crisis of 2008 showed that financial crises can happen even to strong international companies, they must be constantly vigilant about their finances positions [2]. Due to the impact of the global financial crisis, the number of companies going bankrupt rose sharply in many countries, and the importance of financial distress prediction models increased [3], which was of great significance for corporate sustainable development

Objectives

Methods

Findings

Conclusion