Title

Prediction models of corporate financial distress in the Southeast Asian countries

Author Identifiers

Viet Dung Dinh
ORCID:0000-0003-3233-6273

Date of Award

2020

Degree Type

Thesis

Degree Name

Doctor of Philosophy

School

School of Business and Law

First Advisor

Professor Robert Powell

Second Advisor

Dr Duc Vo

Abstract

The development of corporate financial disturbance prediction models plays an essential role in the assessment and management of credit risk. The study will examine two common types of credit risk measurement models: (i) accounting-based models; and (ii) market-based models. Analysis of financial statements is used in accounting-based models to derive a score that shows the differences between distressed and non-distressed firms. Alternatively, market or structural models are developed based on a combination of balance sheet items and volatility in the market values of firms’ assets to measure Distance to Default (DD). These two types of models are applied to the ASEAN region, which has been working towards financial integration across nations within the region. Each ASEAN nation has a very different banking system. This study focuses on the six largest countries in the ASEAN Economic Community (AEC), comprising of Indonesia, Malaysia, the Philippines, Singapore, Thailand, and Vietnam.

A key differentiation in this study is that while many credit risk studies focus mainly on bankruptcy, this study focuses on early warning distress indicators that signal distress well before bankruptcy. This is when firms experience difficulty in servicing debt as measured by interest cover ratio (ICR) and non-performing loans (NPLs). Thus, in this study, default prediction scores are compared to ICR, at a firm level, and to NPLs at a country level.

Multiple Discriminant Analysis (MDA) modelling has been critiqued as having limited ability to predict financial distress if applied to types of firms and economic circumstances which are different to those used in the model’s development. To address these limitations, this study explores various combinations of accounting-based variables in MDA modelling over 720 firms from a broad range of industries and across six countries within the ASEAN region, using both forward-looking and back-testing approaches, over a period of 20 years which includes both the Asian and Global Financial Crises.

The results show that significant accounting-based models can be developed for the countries in the ASEAN region, but there are differences in models from country to country and period to period. Nonetheless, there are some common explanatory variables, particularly those relating to profitability, which dominate across the models. Generally, models developed for specific time periods and countries perform somewhat better than one-size-fits-all models. A higher degree of accuracy is evident when there is less volatility in the sample.

The study also explores the addition of a market-based indicator (DD) to the accountingbased variables. The addition of the market-based DD variable is found to increase the prediction accuracy of the accounting-based variables for only a few countries or time periods.

Key contributions from this study can be summarised as follows. First, the relative advantages and disadvantages of each model are assessed. Second, the effectiveness of each model as an early warning distress indicator over different economic conditions is then compared. Third, the most appropriate model, given the specific circumstances of each country in the research sample, will be recommended to provide evidence-based choices for lenders and regulators in the selected ASEAN nations.

Access Note

Access to this thesis is embargoed until 23 June 2021. At the expiration of the embargo period, access to the thesis will be restricted to current ECU staff and students. Email queries to library@ecu.edu.au

Access to this thesis is restricted. Please see the Access Note below for access details.

Share

Paper Location

 
COinS