Content area
Full Text
ABSTRACT
Nowadays, many banks have their own system to identify good clients from bad clients. Even thou we talk about persons or companies, many banks rely their decision for giving a loan on the loans history from their databases systems, the financial results of the entity (revenue for person and profit for companies), or the "friendship" between the bank and the client (a bank's friend is an old client). Very few banks develop a market analysis or a credit score model in order to analyze the market on which the company operates. In this article, we propose a credit score model, taking into account 12 activity indicators for 21 companies in the real estate market. Using the principal component analysis, cluster analysis and discriminant analysis, we obtain two score functions that can make the difference between companies with a low risk of bankruptcy and companies with a high risk of not refunding the loan.
Keywords: real estate market, credit score, principal components analysis, cluster analysis, discriminant analysis
1. INTRODUCTION
Starting from the Altman (1968) studies about linear discriminant analysis used to predict bankruptcy the recent studies reveal a much more complicate problem when we talk about bankruptcy and models that can predict this fact. The credit score model is very important for a bank's management system, because, using a model like this, a manager can decide whether he gives the loan or not.
The main goals of this article are to identify the variables that fit to a bankruptcy predictable model in the real estate field and to determine the function (discriminant function) that can "decide" if a new firm (that is not analyzed in the model) may get a loan or not according to the real estate market. Finally, we will provide the answer to the questions: "what is the function (or functions) that can help a bank manager to decide giving the loan to a listed company from the real estate field?", "is this model a good model that a manager can rely on?", and "how precise is this model?".
Section 2 of the article presents a part of the existing literature in this domain and some of the newest articles about bankruptcy's prediction. Section 3 reveals the methodology used to...