Bankruptcy prediction for credit risk using neural networks: A survey and new results
Summary (1 min read)
Introduction
- In this study the authors focus only on the corporate bankruptcy prediction problem.
- It can help in accurately assessing thecredit risk of bank loan portfolios.
- The focus of this article is on the empirical approach, especially the use of NNs.
A. Early Empirical Approaches
- Altman uses the classical multivariate discriminant analysis technique (MDA).
- It is based on applying the Bayes classification procedure, under the assumption that the two classes have Gaussian distributions with equal covariance matrices.
- These particular financial ratios have been widely used as inputs, even for NNs and other nonlinear models.
- Ohlson introduced the logistic regression approach (LR) to the bankruptcy prediction problem.
B. Neural-Network (NN) Approaches
- Research studies on using NNs for bankruptcy prediction started in 1990, and are still active now.
- One of the first studies to apply NNs to the bankruptcy prediction problem was the work by Odom and Sharda [33].
- The total assets gives some indication of the size of the firm.
- Boritz and Kennedy [9] (see also [10]) compared between a number of techniques, including different NN training procedures, LR and MDA, using the indicators chosen by Altman, and those chosen by Ohlman.
- Their method obtains better results than NN and MDA for Type II error, but worse results for Type I error.
C. A Brief Review of the Structural Approach
- One of the earlier and commonly used methods is the assetbased approach by Merton [31] (developed further by Longstaff and Schwartz [28]).
- This model views a firm’s equity as an option on the firm (held by the shareholders) to either repay the debt of the firm when it is due, or abandon the firm without paying the obligations.
- This model has been successfully developed into a successful commercial product by KMV Corporation.
- Another approach, by Jarrow and Turnbull [20], models default as a point process, where the time-varying hazard function for each credit class is estimated from the credit spreads.
- J.P. Morgan’s CreditMetrics product [14] is based on modeling changes in the credit quality ratings.
D. Challenges for the NN Prediction Models
- For portfolio credit risk estimation, this is essential in order to compute the loss level.
- As such, indicators obtained from the stock price can be beneficial especially in long horizon default forecast.
- Using an initial prescreening procedure based on individual indicator prediction accuracy and correlation matrix, and then a subsequent cross-validation procedure to narrow down the choice, the authors select the best five or six inputs from this pool of indicators.
- He also received theYoung Investigator Award from the International Neural Network Society, in 1996.
Did you find this useful? Give us your feedback
Citations
978 citations
Cites background or methods from "Bankruptcy prediction for credit ri..."
...[7] Firms USA 1160 BPNN with novel indicators 3 years [8] Firms Finnish 74 DA/DA, LA/LA, GA/BPNN, DA/ BPNN, LA/BPNN 1986–1989...
[...]
...Atiya [7] reviewed the applications of NN in bankruptcy prediction and developed an NN....
[...]
...[1] Cash flow/total liabilities, current assets/current liabilities (current ratio), inventories turnover, net income/total sales, net income/total assets, net working capital/total assets, owners equity/total assets, (total borrowings + bonds payable)/total assets [2] Net income/total assets (return on assets (ROA)), net loan losses/adjusted assets, net loan losses/total loan, (net loan losses + provision for loan losses)/income, non-performing loans/total assets [5] Common equity/total capital (capitalization), cumulative profitability, debt services, stability of earnings, roa, liquidity, size [6] Debt cost, debt quality, growth, indebtedness, share of labour costs, short-term liquidity, size, turnover of assets [7] Book value/total assets, cash flow/total assets, gross operating income/total assets, ROA, price/cash flow, rate of change of cash flow per share (ROC), rate of change of stock price, stock price volatility [8] Cash/current liabilities, cash/net sales, cash/total assets, cash flow/current liabilities, cash flow/total assets, cash flow/total debt, current ratio, current assets/net sales, current assets/total assets, current liabilities/equity, earnings before interest and taxes (EBIT)/total interest payments, equity/fixed assets, equity/net sales, inventory/net sales, long-term debt/equity, market value of equity/book value of debt, net income/total assets, net quick assets/inventory, net sales/total assets, operating income/total assets, quick assets/current liabilities, quick assets/net sales, quick assets/total assets, rate of return/common stock holders, retained earnings/total assets, return on stock, total debt/equity, total debt/total assets, working capital/equity, working capital/net sales, working capital/total assets...
[...]
728 citations
Cites background or methods from "Bankruptcy prediction for credit ri..."
...Keywords: Support vector machines; Bankruptcy prediction...
[...]
...…such as neural networks (NNs) can be an alternative method for classification problems to which traditional statistical method have long been applied (Atiya, 2001; Barniv, Agarwal, & Leach, 1997; Bell, 1997; Boritz & Kennedy, 1995; Charalambous, 0957-4174/$ - see front matter q 2004 Elsevier Ltd....
[...]
...…non-linear relationships in the data set, they have been studied extensively in the fields of financial problems including bankruptcy prediction (Atiya, 2001; Barniv et al., 1997; Bell, 1997; Boritz & Kennedy, 1995; Charalambous et al., 2000; Etheridge & Sriram, 1997; Fletcher & Goss, 1993;…...
[...]
...In this study, we show that the proposed classifier of SVM approach outperforms BPN to the problem of corporate bankruptcy prediction....
[...]
691 citations
479 citations
Cites background or methods from "Bankruptcy prediction for credit ri..."
...Keywords: Bankruptcy prediction; Credit scoring; Neural networks; Classifier ensembles...
[...]
...…intelligence and machine learning techniques (e.g. artificial neural networks (ANN), decision trees (DT), support vector machines (SVM), etc.) have been used to solve the above financial decision-making problems (e.g. Atiya, 2001; Huang, Chen, Hsu, Chen, & Wu, 2004; Lee, Chiu, Chou, & Lu, 2006)....
[...]
...Thus, predicting bankruptcy timely and correctly has become great importance for financial institutions (Atiya, 2001; Zhang, Hu, Patuwo, & Indro, 1999)....
[...]
428 citations
Cites methods from "Bankruptcy prediction for credit ri..."
...Related approaches have been used by Atiya (2001), Shin et al. (2005), and Min et al. (2005), in which artificial neural networks (ANN) and support vector machines (SVM) were applied to the problem of predicting corporate bankruptcies....
[...]
References
13,033 citations
11,225 citations
10,737 citations
5,244 citations
4,210 citations
"Bankruptcy prediction for credit ri..." refers background in this paper
...The pioneers of the empirical approach are Beaver [7],...
[...]
...The pioneers of the empirical approach are Beaver [7], Altman [2], and Ohlson [34]....
[...]
...Beaver was one of the first researchers to study the prediction of bankruptcy using financial statement data....
[...]
Related Papers (5)
Frequently Asked Questions (11)
Q2. What are the main indicators for the financial ratio and equity-based system?
For the financial ratio and equity-based system the chosen indicators are:1) book value/total assets BV/TA; 2) cashflow/total assets CF/TA; 3) price/cashflow ratio P/CF; 4) rate of change of stock price ROC(P); 5) rate of change of cashflow per share ROC(CF); 6) stock price volatility VOL.
Q3. What is the next step for the research community?
Once this is established, the logical next step for the research community is to improve further the performance of NNs for this application, perhaps through better training methods, better architecture selection, or better inputs.
Q4. How much accuracy did Altman et al. achieve for the financial ratio system?
For the financial ratio system the authors obtained a prediction accuracy of 84.52% for the in-sample set, and 81.46% for the out-of-sample set.
Q5. What is the way to estimate the probability of default?
Wilson [54], [55] proposed a discrete-time dynamical model, whereby the default probabilities are a function of macro-economic variables.
Q6. What is the main idea of the paper?
In this work the authors introduce a novel set of indicators that can be used in addition to the financial ratios and lead to significant improvement in prediction accuracy.
Q7. How accurate is the financial ratio system?
For the financial ratio and equity-based system the authors obtained a prediction accuracy of 89.41% for the in-sample set, and 85.50% for the out-of-sample set.
Q8. What is the method for predicting bankruptcies?
Fan and Palaniswami [59] propose the use of support vector machines (SVMs) for predicting bankruptcies among Australian firms, and compared it with NN, MDA and learning vector quantization (LVQ).
Q9. How many inputs are selected from the NN?
Using an initial prescreening procedure based on individual indicator prediction accuracy and correlation matrix, and then a subsequent cross-validation procedure to narrow down the choice, the authors select the best five or six inputs from this pool of indicators.
Q10. How did they achieve the accuracy of NN?
They applied their technique to a collection of Belgian firms (no forecast horizon is mentioned), where they obtained an accuracy of 82.9% versus 76.1% for the nontransformed input case.
Q11. What is the corresponding acuracy for type II?
The corresponding results for MDA were in the range of 59.3% to 70.4% for Type The authoracuracy, and in the range of 78.6% to 85.7% for Type II acuracy.