Advanced Analytical Models for Financial Accuracy and Risk Stratification

Despite the increasing use of sophisticated models by financial institutions, the financial crisis of 2008–09 has raised serious doubts about their predictive reliability, especially in the area of risk assessment and capital allocation. New tools are needed to build and validate models of financial accuracy, risk, and credit for all market participants. Risk is a measure of the possibility of adverse future outcomes, but a surrogate measure called credit risk score is widely used for operational convenience. A rigorous definition of accuracy is proposed, along with data provenance and quality assurance procedures. The important sources of inaccuracy are identified, and broad modeling frameworks for financial accuracy, risk stratification, and risk controlled return optimization are outlined.