Using data mining and intelligence for predictive behavior analysis can stop potential losses from mortgage fraud. Today the technology, the analytics and the processes exist to deter bad acts and to uncover bad actors BEFORE they commit a crime that can cost you and your customers tens of thousands of dollars in actual fraud losses.
Although you’ve probably heard many times that predictive analytics (like those employed by Google© and other high tech marketing firms) will optimize your marketing campaigns, it’s hard to envision, in more concrete terms, what it will do with risk management. How can you get a handle on the functional value of data mining, analysis and reporting as a tool for originating and selling higher quality loans less likely to default or worse, less likely to be a loan file rife with mortgage fraud? The answer lies in knowing what to look for.
Predictive analytics rely upon certain key factors, the central building block being the predictors, which are values ascribed to the type of behavior you wish to predict. In the case of mortgage fraud the predictors can include: years of experience; existence of a license and insurance; business and professional history (i.e. litigation and regulatory discipline); liens and judgments. These and other factors can give a comprehensive picture of the potential for a professional to cut corners, make mistakes, engage in self-dealing, and be open to participating in conspiracies to commit fraud in return for financial rewards.
There are several companies in the mortgage space selling tools implying predictive behavior analytics, however not all of them engage in the complex analytical analysis and reporting that provides true value for deterring and detecting fraud. This may be why a decade after so much loan origination fraud tool technology entered the marketplace fraud is increasing, not decreasing.
It is not enough to simply gather such data and to spit it out for general consumption. The real value to the data mining that results in collecting predictors is the risk analysis that takes place, some of which may be accomplished through automation but still requires human evaluation. The trick is to find the best predictive model.
At SSI, we built a system that used fraud statistics from the past decade to build a predictive model specialized for the mortgage industry. The process learns from the data’s collective values. It also involves not just automation but trained vetting specialists who review key data points to discount false positives, provide more insight into derogatory public data, and thereby provide more reliable risk ratings for the agents. At the same time it offers a fair and reasonable process for the agents themselves, who may appeal negative findings and work with SSI staff to clear misleading data that emanates from sometimes unreliable public data sources.
The most immediate value to subscribing to a risk analytics tool for your business is the deterrent factor. Before SSI launched, it conducted two years of beta testing with large and small lenders nationwide. During that time we found a 23% high risk knock-out rate among approved and active agents on lender lists. Since SSI began vetting the high risk rate is 2%. Quite frankly when bad actors know they are being subjected to a comprehensive data mining and intelligence process, they stay away, far away. And that is good news for the mortgage industry.