At the 2011 SAS Global Forum Executive conference, we were treated to an interesting discussion on using analytics to detect fraud. The discussion (moderated by our own Stewart Bradley, Senior Business Director, Financial Crimes and Fraud) spanned two business areas impacted by fraud and included:
- Manuel Moreno, PhD, Director of Research and Evaluation, Los Angeles County Chief Executive Office: Manuel’s analytic team covers fraud analytics around social services provided by Los Angeles County – one of the largest and most diverse counties in the United States.
- James Schweitzer, SVP and COO, National Insurance Crime Bureau (NICB): The NICB is a non-profit dedicated to tackling fraud for their member companies, typically law enforcement bureaus and property-casualty insurance companies. Member companies submit claims data to NICB, which can then aggregate the information and look for cross-industry trends in fraud that an individual member company may not see.
Combined, the dollar impact of fraud for these organizations reaches hundreds of millions of dollars a year. In fact, the impact is so big that standard methods and techniques of investigating fraud were becoming less effective: The investigative processes were manual and cumbersome, and often investigators had to chase down a potential fraud case without having a good leading indicator that the case was worth spending time on. Most companies don’t have unlimited investigative resources: Higher tech tools and techniques are essential in helping identify and prioritize potential fraud to maximize those resources time and cost. For these organizations, the key was addressing fraud through applied analytics and specifically leveraging SAS’s Fraud Framework. What have these organizations been challenged with? What is making their initiatives successful? As we listen to the discussion, a few common themes emerge:
- The need for an automated approach:
- “Manual efforts are not efficient: Investigations are time consuming and expensive. When we were faced with the fiscal crisis starting in 2007, we needed to put more discipline in the process.”
- “We’re responsible for 175 investigators in the US – many of the came from law enforcement and the insurance industry. Our job is to convince prosecutors to take cases to trial: We need a compelling story and documentation to present to law enforcement and prosecutors.”
- It’s all about getting the data together…and also finding new opportunities in the data:
- The NICB has been able to expand the scope of its mission: “We’re able to look across our 'all-claims database' – our organization sits below the top tier of the information, which means that we can look across the industry. We aggregate claims to look for organized groups that are committing fraud…Our focus has traditionally been on [fraud related to motor-] vehicle crimes, but over the last 5-10 years, we’ve changed our focus to medical fraud. We’re seeing numerous schemes trying to grab the medical components of property-casualty insurance, specifically in auto, but also workers’ compensation and health insurance.” In expanding their capabilities with the fraud detection, strategic analysts are putting together forecast reports that can identify different types of fraud, fraud rings and provider fraud. “We can send those [predictive] leads to our member companies for further investigation.” In addition, the NICB is looking at fraud detection in billing (the current detection work is done on claims). “We can take the billing data and apply the fraud algorithms and will be rolling out alerts in the next couple of weeks. We believe this will be a tremendous value to our partners. We’re reaching out to the healthcare industry because there are multiple opportunities to commit fraud – it’s bigger than just the property-casualty insurance industry. There’s very little data sharing taking place between different types of providers.”
- Culture change is a critical success factor:
- Stewart asked the panel: What cultural barriers have you run into and how did you overcome them? For the NICB, change management was essential. “Many of our investigators are seasoned professionals: They need to be educated on the importance of the new fraud detection tools. The investigators are used to using traditional skills, such as surveillance or interviews, and those are valuable, but when you start to add data, it enhances their ability to investigate more effectively. Our investigators are evaluated on their ability to get cases prosecuted. White collar crime can be difficult and time consuming to prosecute. We need to package the investigation in a way that makes it easy for us to hand it over to law enforcement and prosecutors, where most of the work has been done. By leveraging analytics, we can make the process easy for prosecutors and explainable to a jury.”
- For Los Angeles County in moving from a manual to an automated environment, “We had a mandate to look at the feasibility of implementing data mining for fraud detection programs; the attitude among many was that it was a political initiative – nothing will happen; when we began the proof of concept, our investigators were concerned about the workload change – investigators had too many fraud areas that they were investigating and they were overloaded. We worked for six months with investigators and administrators to understand the deficiencies and they began to feel more comfortable with changes and technology. Any time you change the way people work, you have discomfort; at the end of our pilot, the results were generally positive. We were able to convince the investigators that the detection techniques would help prioritize the right opportunities to investigate and free them from pursuing cases that weren’t relevant.”
Stewart wrapped up the session with some parting thoughts: He’s seeing a trend in the sharing of data across industry and organizational lines; utilizing portfolio of analytics to solve problems; and taking an enterprise approach to combating fraud at an industry level. As illustrated by our panel, tackling fraud can be a daunting and rewarding challenge, but analytics helps uncover hidden relationships, prioritize opportunity and make your investigative resources more effective.