Developing a Neural Network Model to Assign an Implied S&P Credit Rating

Prior to entering into a contract with a company where one may be financially exposed, it is important to access the company’s financial health. Specifically, it is imperative to determine the company’s capacity to meet its financial obligations to pay off debt. Gratefully, the credit risk industry has developed hierarchical credit rating systems to inform the public of the risk of default, relative to other public companies, based on their financial metrics. The three largest credit rating agencies, Standard & Poor’s, Moody's and Fitch Group, rate a large portion of companies who publicly trade their stock. However, there are tens of thousands of private companies who are not rated by these agencies. Many of these companies are willing to provide their audited financial metrics (confidentially) to a potential creditor or customer so they can to determine their credit worthiness. This webinar demonstrates how to build a neural network model using NeuralTools from the DecisionTools Suite of @RISK, that will assign an implied S&P credit rating to a private company who provides the necessary financial metrics. The webinar will discuss where to acquire the metrics for thousands of companies needed to build the model, how to scrub the financial data, how to select the final set of metrics to be used in the model, how to determine the size of the test group, how to identify and correct for possible bias, and how to apply a validation process to select the final neural network model.