DESIGNING AN EXPERT SYSTEM FOR COMMERCIAL LOAN DECISIONS by Huynh Ngoc Lieu A Research Study submitted in partial fulfillment of the requirements for the degree of Master of Business Administration Examination Committee Prof. Jyoti P. Gupta (Chairman) Dr. Do Ba Khang Dr. R. Sadananda Nationality Vietnamese Previous degree Bachelor of Economics HoChiMinh City University of Economics HoChiMinh City, Vietnam Scholarship Donor Government of Switzerland Asian Institute of Technology School of Management Bangkok, Thailand April, 1996 ACKNOWLEDGEMENTS Behind every achievement lies an unfathomable sea of gratitude to those who actuated it, without whom it would never have come into existence. To them I lay the words of gratitude imprinted deep within me. I place on record my gratitude to my Advisor, Prof. J.P.Gupta, who kindly gave me the valuable time, knowledge, and advice to develop and complete my research. I owe a deep indebtedness, which I acknowledge with pleasure and am sincerely thankful to my Committee members, Dr Do Ba Khang and Dr. R. Sadananda, for their precious advises in serving as members of the examination committee. Thanks are due to Mr. Dominique Rohner, Director of SAV program for his help and encouragement during the learning period in Vietnam as well as AIT. I thank all the members of the SAV program, specially those who sincerely gave me a lots of help. I would like to express my deep appreciation to the Government of Switzerland for providing me with a scholarship to pursue my M.B.A at AIT by the SAV program in Vietnam. I wish to thank my all friends, colleagues at my faculty at University of Economics Ho Chi Minh City, specially those who geared up my confidence and revived my spirits, every time I was feeling down. I thank my mother, for giving me everything in my life and teaching me the way it should be. ABSTRACT The primary job of a loan officer is to decide on the conditions of a loan. To do this, the loan officer has to track the applicant’s credit history as well as check previous and current financial status. The nature of the task is repetitive, yet unstructured. To improve the throughput and accuracy of loans granted and to insure greater consistency of loan review, several credit analysis expert system have been developed. This study gives the design of an expert system for commercial loan decisions. It can help loan officers to reduce the time devoted to the analysis in evaluating loan applicants and to improve the quality of the evaluation. The domain of credit analysis application covered by this study is to analize commercial loans for industrial or retail borrowers, and it will work for loans of any size. TABLE OF CONTENTS CHAPTER TITLE PAGE Title Page i Acknowledgments ii Abstract iii Table of Contents iv Abbreviations v List of Figures vi 1 Introduction 1 1.1. Rationale 1 1.2. Objective 2 1.3. Scope of the Research 3 2 Literature Review 4 2.1. Financial Perspective 4 2.2. Existent Systems 8 2.3. Relative Comparison between Methodologies 9 3 Theoretical Considerations 12 3.1. Credit Analysis: What Make a Good Loan? 12 3.2. Financial Ratio Analysis of a Customer’s Financial Statements 14 3.3. Commercial Loan Decision Making Process 19 3.4. Architecture of Expert Systems 19 3.5. Applying Expert System to Commercial Loan Decisions 24 4 Design and Methodology 26 4.1. Methodology 26 4.2. Design 26 4.2.1. Overall Structure Design 26 4.2.2. Implementing Knowledge Base Design 28 5 Conclusions and Recommendations 50 5.1. Overview 50 5.2. Recommendations 51 References 53 Appendix 1: Running the Program 55 Appendix 2: System Information 76 Appendix 3: Sample Reports 89 ABBREVIATIONS CP Current Period PP Prior Period PY Prior Year LFY Last Fiscal Year CM Current Maturities LTD Long-Term Debt ELTD Existing Long-Term Debt CMLTD Current Maturities Long-Term Debt NLTD New Long-Term Debt CMNLTD Current Maturities New Long-Term Debt TNW Tangible Net Worth ST Short-Term STD Short-Term Debt N/A Not Applicable SG&A Sales, General and Administrative Expenses COGS Cost of Goods Sold NP Notes Payable Ave Average Pro Prorated CO Class Override Abs Absolute Value B/S Balance Sheet I/S Income Statement R/E Retained Earnings CVLI Cash Value Life Insurance A/P Accounts Payable A/R Accounts Receivable NR Notes Receivable Liab Liabilities Non-Int Non-Interest Non-Curr Non-Current EBITDA Earnings Before Interest Taxes Depreciation Amortization EBIT Earnings Before Interest Taxes Sub Subordinated LIST OF FIGURES FIGURE CAPTION PAGE 2.1 Loan Negotiation Phases 4 2.2 Hierarchical Model of the Loan Evaluation process 6 3.1 Commercial Loan Decision Making Process 20 3.2 Generic Architecture of Conventional Expert System 21 4.1 Overall Structure Design 27 4.2 Database Structure 28 4.3 Rulebase Structure 44