Doomed Firms : An Econometric Analysis of the Path to Failure

"This title was first published in 2003. This book provides a much-needed comprehensive and up-to-date treatise on financial distress modelling. Since many of the challenges facing researchers of financial distress can only be addressed by a totally new research design and modelling methodology, this book concentrates on extending the potential for bankruptcy analysis from single-equation modelling to multi-equation analysis. Essentially, the work provides an innovative new approach by comparing each firm with itself over time rather than testing specific hypotheses or improving predictive and classificatory accuracy. Added to this new design, a whole new methodology - or way of modelling the process - is applied in the form of a family of models of which the traditional single equation logit or MDA models is just a special case. Preliminary two-equation and three-equation models are presented and tested in the final chapters as a taste of things to come. The groundwork for a full treatise on these sorts of multi-equation systems is laid for further study - this family of models could be used as a basis for more specific applications to different industries and to test hypotheses concerning influential variables to bankruptcy risk."--Provided by publisher Cover Half Title Dedication Title Page Copyright Page Contents List of Figures List of Tables Preface PART I: THEORETICAL ISSUES 1 Introduction 1.1 The Distress Continuum 1.2 Study Methodology 1.3 Background and Position of the Book 1.4 Structure of the Book 1.5 Publications 2 Precursors and Challenges in the Literature 2.0 Introduction 2.1 The Pioneering Models 2.1.1 Beaver (1966) - Univariate Models 2.1.2 Altman (1968) and Early Followers - Multiple Discriminant Analysis Models 2.2 The Research of the 1980s – The Logit Model 2.2.1 Ohlson (1980) 2.2.2 Zavgren (1983) 2.3 The Choice of Independent Variables and Theories of Bankruptcy 2.3.1 Models from the Probability and Statistics Literature 2.3.2 The Cash Flow Model 2.4 The Time Dimension and Macro-Economic Conditions 2.4.1 Rose, Andrews and Giroux (1982) 2.4.2 Mensah (1984) 2.4.3 Kane, Richardson and Graybeal (1996) 2.5 Statistical Methodology 2.5.1 Zmijewski (1984) 2.5.2 Marais, Patell and Wolfson (1984) 2.6 The Distress Continuum 2.6.1 Survival Analysis 2.6.2 Formal Models 2.6.3 Gilbert, Menon and Schwartz (1990) 2.6.4 Flagg, Giroux and Wiggins (1991) – A Failing Company Model 2.6.5 Multivariate Time Series Process – Theodossiou (1993) 2.6.6 Hill, Perry and Andes (1996) 2.7 Emerging Technologies 2.7.1 Neural Networks 2.7.2 Chaos Theory 2.7.3 Latest Techniques 2.8 Major Challenges and Future Directions 2.9 Conclusion 3 Failure of Firms as Process: A New Modelling Methodology 3.0 Introduction 3.1 Problems in Empirical Single-Equation Bankruptcy Analysis 3.1.1 Estimating Probabilities of Failure 3.1.2 A Critical View 3.2 Emergent Research Objectives 3.2.1 Valid Accommodation of the Available Data into Appropriate Models 3.2.2 Addressing the Problem of Sample Selection Bias 3.2.3 Improving Model Specification for Relevance over Time 3.2.4 Improving Model Specification to Reflect Dynamics 3.2.5 Improving Model Completeness 3.3 A New Methodology 3.4 Constructing and Testing General Multi-Equation Models 3.4.1 Techniques for Estimating the Parameters in a Linear Simultaneous System of Equations 3.5 Applying Multi-Equation Models to the Area of Financial Distress 3.5.1 A Family of Models for Estimating the Risk of Failure 3.6 Conclusion PART II: EMPIRICAL ISSUES: TOWARDS A DYNAMIC MODELLING OF FINANCIAL DISTRESS 4 The Internal Environment: The Financial Ratios 4.0 Introduction 4.1 The Relationship Between the Financial Variables and Financial Distress 4.1.1 Modelling Bankruptcy 4.2 The Ratios 4.3 The Data 4.3.1 Missing Values 4.3.2 Measurement Problems: Extreme Values and Confounded Ratios 4.3.3 Ratios where Measurement Problems Existed 4.3.4 Descriptive Statistics: Mean Values For The Ratios by Lag Year 4.4 The Binary Dependent or Outcome Variable – A Definition of Failure 4.5 A Logit Analysis of Bankruptcy Risk Related to the Internal Variables of the Firm 4.5.1 Stepwise Analysis in Stages 4.5.2 The Significant Financial Variables to Failure Risk 4.5.3 A Note on Classification 4.5.4 The Signs of the Coefficients 4.6 Conclusion 5 The External Environment: The Impact of the Macroeconomy 5.0 Introduction 5.0.1 The Failing Firm in the Macroeconomy 5.0.2 Some Methodological Issues 5.1 Principal Components Analysis of the External Economic Variables of the USA 5.1.1 The Five External Factors 5.2 Statistical Artifact of the ‘Economy Growth’ Factor 5.3 Structuring the Variables Representing the External Environment: Distributed Lag Models on the Principal Components 5.3.1 Experimentation with Different Lag Models Using the Service Industry Dataset 5.3.2 Conclusions on the Lag Model Weights 5.4 The External Variables Logit Model 5.4.1 The Meaning of the External Variables Model 5.5 Conclusions 6 Combining the Impact of the Internal and External Factors on Financial Distress 6.0 Introduction 6.1 A Combined Financial Distress Model for Bankrupt Service Companies 6.1.1 The Significant Financial Variables to Failure Risk 6.1.2 The Significant External Variables to Failure Risk 6.1.3 The Combined Model Equation Assuming Independence Within Firm-Years 6.1.4 The Sign of the Coefficients 6.1.5 Independence of the Internal and External Variable Sets on Failure Risk 6.2 How Important are the Economic Variables Compared to the Financial Variables in Influencing Failure Risk? 6.3 Comparison of the Average Estimated Probabilities of Failure for Each Lag Year 6.4 Applying the Failure Risk Model to Another Industry 6.5 Conclusion 7 A Dynamic Model of Financial Distress: The Path to Failure 7.0 Introduction 7.1 Inclusion of Lagged Risk in a Single-Equation Logit Model of Failure Risk 7.1.1 Sensitivity Analysis to Failure Risk Estimated for the Fifth Year Prior to Failure 7.1.2 Successive Model Iterations of Failure Risk on Lagged Risk 7.2 Multi-equation Models 7.2.1 Seemingly Unrelated Regression Models 7.2.2 Simultaneous Equation Models 7.2.3 Working Capital/Total Assets as a Dependent Variable 7.2.4 Interest Coverage after Tax as a Dependent Variable 7.3 Model Dynamics 7.4 Conclusions PART III: DISCUSSION AND CONCLUSIONS 8 Conclusion 8.0 Introduction: Towards Methodological Cohesion 8.1 Departures from the Past 8.2 Empirical Results 8.3 Limitations 8.4 Implications of the Research and Possible New Directions 8.5 Concluding Comments PART IV: APPENDICES Appendix A—Appendix P Bibliography Index