I recently attended an engaging conference jointly hosted by The Risk Management Institute and National University of Singapore that explored “risk management responses to rising systematic and systemic risks.” Among its distinguished speakers were various local and international banks, professors from NUS and National Taiwan University, the Federal Reserve, IMF and the Monetary Authority of Singapore. From the discussions, one issue was exceptionally clear – understanding models, and whether they work or don’t, is increasingly important in a financial world that has become so volatile.
Predictive accuracy and model reliability have come under tremendous regulatory scrutiny. Adding to this challenge is the fact that data-driven tools become less and less reliable as time goes by. Because of this, regular and effective model validations have become critical.
According to US regulators, models can improve business decisions, but they also impose risk, including the potential for adverse consequences from decisions based on models that are either incorrect or misused. Rumor has it that US regulators are turning their attention to marketing models as well, which was not the case a few years ago. For large organizations, especially those operating in multiple countries, this may mean monitoring hundreds of models on an ongoing basis.
Fortunately, new tools are helping to automate this type of model management housekeeping. One benefit, among many, is that validation teams are freed up to use their expertise on strategic decisions, rather than generating routine MIS to figure out whether anything is amiss, which is currently the norm at many organizations. Indeed that’s our goal behind FICO® Model Central™ Solution: to remove the routine high-pressure “firefighting” often involved in managing predictive models, so our clients can instead focus on making the right decisions to grow their business.
