Model Risk: The Challenges of Legacy Code and Best Practices

The model risk inherent in derivative pricing models has received a burst of renewed interest since the global financial crisis. Driven partly by regulators and partly by internal pressure to avoid losses due to poor modeling, the emerging trend is for financial institutions to utilize techniques for measuring and mitigating model risk. Although these ideas have existed in some form for many years, this revived attention to the field has sparked a renewed inquiry into best practices for development of pricing models. In this paper, we will examine the most common types and sources of model risk, and then outline best practices that practitioners can utilize in their model validation processes.