We address a number of comparative issues relating to the performance of failure prediction models for small, private firms. We use two models provided by vendors, a model developed by the National Bank of Belgium, and the Altman Z-score model to investigate model power, the extent of disagreement between models in the ranking of firms, and the design of internal rating systems. We also examine the potential gains from combining the output of multiple models. We find that the power of all four models in predicting bankruptcies is very good at the one-year horizon, even though not all of the models were developed using bankruptcy data and the models use different statistical methodologies. Disagreements in firm rankings are nevertheless significant across models, and model choice will have an impact on loan pricing and origination decisions. We find that it is possible to realize important gains from combining models with similar power. In addition, we show that it can also be beneficial to combine a weaker model with a stronger one if disagreements across models with respect to failing firms are high enough. Finally, the number of classes in an internal rating system appears to be more important than the distribution of borrowers across classes.

Failure prediction models: performance, disagreements, and internal rating systems. NBB Working Papers. No. 123, 13 December 2007

Purpose – The aim of this paper is to develop a bankruptcy prediction model for the Belgian small- and medium-sized enterprises (SMEs) through the building of a logit model that includes a selection of financial ratios. Design/methodology/approach – Using a sample of 7,152 Belgian SMEs among which 3,576 were declared bankrupt between 2002 and 2012, the model, which includes control variables such as firm size and age, aims to test the predictive power of ratios reflecting the financial structure, the profitability, the solvency and the liquidity of firms. Findings – The results report a satisfactory prediction accuracy and show that ratios as profitability and liquidity are excellent predictors of bankruptcy for Belgian SMEs. Research limitations/implications – Although the results seem to be conclusive, it could be noted that the healthy sample was not paired with the bankrupt sample. Other studies show that the use of paired samples makes it possible to increase the already good prediction rate. Also, f...

Bankruptcy Prediction: The Case of Belgian SMEs

Purpose – The objective of this paper is to stress the importance of detecting financial frauds in predicting business failures disclosed by the unexpected financial crisis brought by Enron, Worldcom and other corporate distresses involving accounting irregularities.Design/methodology/approach – The most frequently used methodologies in predicting business failures, discriminant analysis and neural network (NN) (based on the Kolmogorov‐Gabor polynomial Volterra series algorithm) are used. This paper suggests a two‐stage NN procedure: the first stage detected the false financial statements, which were excluded from samples that used to predict the business failures at the second stage. The one‐stage discriminant analysis and the NN model are used to contrast the two‐stage approach in terms of accuracy rate.Findings – The one‐stage NN model has a higher accuracy rate in identifying failed firms than the discriminant analysis, while the two‐stage NN approach has an even higher accuracy rate than the one‐stag...

Predicting business failure under the existence of fraudulent financial reporting

Belgium has faced an important number of corporate bankruptcies during the last decade. The aim of this paper is to develop a model that predicts bankruptcy using three financial ratios that are simple and easily available, even for small businesses. We used a sample of 3,728 Belgian Small and Medium Enterprises (SME’s) including 1,864 businesses having been declared bankrupt between 2002 and 2012 and conducted a neural network analysis. Our results indicate that the neural network methodology based on three financial ratios that are simple and easily available as explanatory variables shows a good classification rate of more or less 80 percent. Results of this study may be of interest for financial institutions and for academics.

/pdf/bankruptcy-prediction-model-using-neural-networks-31fhznbbrl.pdf

Bankruptcy Prediction Model Using Neural Networks

We investigate the relevance of aggregate and consumer-specific income uncertainty for aggregate consumption changes in the US over the period 1952-2001. Theoretically, the effect of income risk on consumption changes is decomposed into an aggregate and into a consumer-specific part. Empirically, aggregate risk is modelled through a GARCH process on aggregate income shocks and individual risk is modelled as an unobserved component and obtained through Kalman filtering. Our results suggest that aggregate income risk explains a negligible fraction of the variance of aggregate consumption changes. A more important part of aggregate consumption changes is explained by the unobserved component. The interpretation of this component as reflecting consumer-specific income risk is supported by the finding that it is negatively affected by received consumer transfers.

/pdf/income-uncertainty-and-aggregate-consumption-kaygjiiy0p.pdf

Income Uncertainty and Aggregate Consumption

We give an overview of the shortcomings of the most frequently used statistical techniques in failure prediction modelling The statistical procedures that underpin the selection of variables and the determination of coefficients often lead to ‘overfitting’ We also see that the ‘expected signs’ of variables are sometimes neglected and that an underlying theoretical framework mostly does not exist Based on the current knowledge of failing firms, we construct a new type of failure prediction models, namely ‘simple-intuitive models’ In these models, eight variables are first logit-transformed and then equally weighted These models are tested on two broad validation samples (1 year prior to failure and 3 years prior to failure) of Belgian companies The performance results of the best simple-intuitive model are comparable to those of less transparent and more complex statistical models

Business failure prediction: simple-intuitive models versus statistical models

This paper gives an overview of the shortcomings of the most frequently used statistical techniques in failure prediction modeling. The statistical procedures that underpin the selection of variables and the determination of coefficients often lead to ‘overfitting’. It is also seen that the ‘expected signs’ of variables are sometimes neglected and that an underlying theoretical framework mostly does not exist. Based on the current knowledge of failing firms, this paper constructs a new type of failure prediction model, namely ‘Simple-Intuitive Models’ (SIM). In these models, eight variables are first logit-transformed and then equally weighted. These models are tested on two broad validation samples [one year prior to failure (ypf) and three ypf] of Belgian companies. The performance results of the best simple-intuitive model is comparable to those of less transparent and more complex statistical models.

Business Failure Prediction: Simple-Intuitive Models versus Statistical Models

We give an overview of the shortcomings of the most frequently used statistical techniques in failure prediction modelling. The statistical procedures that underpin the selection of variables and the determination of coefficients often lead to ‘overfitting’. We also see that the ‘expected signs’ of variables are sometimes neglected and that an underlying theoretical framework mostly does not exist. Based on the current knowledge of failing firms, we construct a new type of failure prediction models, namely ‘simple-intuitive models’. In these models, eight variables are first logit-transformed and then equally weighted. These models are tested on two broad validation samples (1 year prior to failure and 3 years prior to failure) of Belgian companies. The performance results of the best simple-intuitive model are comparable to those of less transparent and more complex statistical models.

Pieter Vandermoere

Papers

Business failure prediction: simple-intuitive models versus statistical models

Business Failure Prediction: Simple-Intuitive Models versus Statistical Models

Business failure prediction: simple - intuitive models versus statistical models