Blog

Pondera FraudCast Blog

rss

Pondera FraudCast

Welcome to the Pondera FraudCast, a weekly blog where we post information on fraud trends, lessons learned from client engagements, and observations from our investigators in the field. We hope you’ll check back often to stay current with our efforts to combat fraud, waste, and abuse in large government programs.

The Irreplaceable Human Mind

The Irreplaceable Human Mind

At Pondera, we are often asked whether fraud detection algorithms will ever completely replace human investigators. And while I can’t address the “ever” part of the question, I can confidently state that it will not happen in the foreseeable future. One of the major reasons for this? Prediction models, like many people, struggle to distinguish between cause and effect.

A Stanford University professor recently shared her studies on this topic which support many of our own findings. She noted that while prediction algorithms are excellent at finding patterns in large data sets, their effectiveness is limited because they struggle with determining causation. An example she used is that algorithms have been shown to help identify patients who should not receive hip surgery because they would likely die of other causes. However, the algorithms are unable to prioritize those patients who should receive the surgery.

In several cases, the professor notes that correlation can be as low as 50%. And she properly notes that while this may be fine in certain situations, governments simply cannot conduct such high-risk experiments with social welfare, economic policies, and other important matters. And unlike controlled environments, such as those that use placebos to test medications, the real world is simply too messy and unpredictable to control all factors.

This problem of causation identifies an important intersection between human reasoning and prediction algorithms. We believe that in complex, rapidly changing environments like fraud detection, effective detection systems combine the power of modern detection algorithms with experienced human reasoning.

By leveraging the individual strengths of both machine and human learning, we can analyze massive data sets and make sense of the findings. We regularly use the system to find the problem and ask the human experts to help explain the problem. This makes the results actionable, which ultimately is what our government partners require.
Ugly Case of Health Care Fraud

Ugly Case of Health Care Fraud

A few weeks ago, I published a blog post titled “Money Obtained Fraudulently is Rarely Used for Good Purposes”. In it, I made the argument that government fraud is a serious, and at times very ugly problem. Now I no longer have to make that argument because the United States Justice Department is making the argument for me.

Last week, the Justice Department announced the largest health care fraud case it’s ever prosecuted; one that defrauded over $1 billion over the past 14 years. The alleged perpetrators of the fraud are said to have leased private jets and chauffeured limousines. One even bought a $600,000 watch! Remember, this is your tax money we’re talking about. The system ran on a complex network of bribes and kickbacks.

And if that’s not enough, here is one of the schemes they allegedly ran. They “treated” seemingly healthy, elderly people with medications they did not need in order to create addictions which would lead to further treatments. Pure evil. Unfortunately, fraudsters are most active where large amounts of money meet vulnerable populations. This is yet another example of that and more reason for us to do what we do.
The False Positive Problem

The False Positive Problem

Last year, 60 Minutes did a segment on the impact of errors in the Social Security Administration’s Master Death File—a database that stores dates of death for Americans. The system stores 86 million records, and despite all best efforts, it still has some issues.

60 Minutes pointed out that errors in the system contribute to millions of dollars in improper payments each year. After all, the system would seem to indicate over 6.5 million Americans over the age of 111 when, in fact, there are probably fewer than 100. On the other hand, false positives where people find themselves mistakenly placed on the list, lead to nightmarish scenarios for obtaining loans, opening bank accounts, and other everyday tasks.

This story demonstrates one of the largest challenges for government agencies: how to use imperfect data sources to minimize fraud, waste, and abuse while also not “harassing” legitimate people and businesses. And unlike a private business that may view a false positive as an inconvenience (who hasn’t had to call their credit card company to say that “yes, that large ice cream purchase was legitimate”), government officials are severely criticized when they act on false positives. In effect, they are criticized for not acting and they are criticized for acting.

Pondera suggests that governments mitigate the effects of false positives by using composite indicators that draw information from multiple sources—both simple data matches like the Master Death File and more complex behavioral sources. For example, a Medicaid investigator would feel much more confident looking into a person who not only shows up on the Master Death File, but also appears to be traveling 100 miles for 20-minute doctor appointments, receives highly unusual (and expensive) procedures for their apparent diagnosis, and often sees two doctors in distant cities on the same day.

Anyone who has worked for or with government program integrity units understands the unique pressures they face. Combining available data sources with intelligent analytics can go a long way toward helping them investigate the right cases while not interfering with program delivery.
Money Obtained Fraudulently is Rarely Used for Good Purposes

Money Obtained Fraudulently is Rarely Used for Good Purposes

People often ask me if I think we can make a difference fighting fraud by stopping down-on-their-luck Americans from grabbing a few extra bucks that they are not entitled to from government programs. In fact, many people ask if it’s even the right thing to do. After all, they explain, wouldn’t anyone do the same given the circumstances?

This illustrates the common misperception that fraud is only perpetrated in small amounts by desperate people who are temporarily bending the rules. In fact, much of what we see takes place on a larger scale. And more importantly, the truth is that money obtained fraudulently is rarely used for good purposes. Examples include:

  • During a “National Counter Terrorism-Awareness Week” in 2014, government officials explained that taxpayer money was being defrauded out of government programs (including student loans) to fund terrorism. In effect, we are helping to fund groups that want to do us harm.
  • The Wall Street Journal reported on, in March 2016, the growing trend of street gangs funding activities through fraud. Fraud offers attractive forms of theft because “they are more lucrative, harder to detect and carry lighter prison sentences”.

Considering that the government distributes over two trillion dollars per year in subsidies, and considering how fraudulently obtained money can be used, it is critical that we address the issue of fraud, waste, and abuse. So to answer the questions posed earlier: Yes, I do believe we are doing the right thing, and Yes, I do believe we can make a difference.

The Internet Changes Everything

The Internet Changes Everything

Way back in 2006, I read an article in the Harvard Business Review that described how the Internet had changed the sales profession. One key observation dealt with the “de-coupling” of the sales cycle from the buying cycle. Prior to the Internet, buyers had to contact vendors for information on their products. Today, buyers do their own research and successful salespeople need to unhinge preexisting customer assumptions prior to starting their sales process.

I believe that the Internet has had an even greater impact on fraud in government benefit programs. Government agencies are under constant pressure to move applications, certifications, and other processes on line to make them more convenient for citizens and businesses. This makes perfect sense because, after all, government exists to serve the needs of the citizens. Unfortunately, moving these processes to the Internet dramatically increases the incidence of fraud.

The Internet provides a degree of anonymity that makes it extremely attractive to fraudsters. The number of fictitious businesses and “ghost beneficiaries” in government programs has exploded in recent years. Many of our customers deal with applications associated with out-of-state or out-of-country IP addresses. Others come from deceased or incarcerated individuals. Still others show indicators of originating in “sweat shops” that create bulk applications and claims.

Just like the salesman that had to adjust to the new sales cycle, it’s important that government program integrity staff adjust to the changing fraud landscape. IP spoofing, anonymous email services, and the wide availability of stolen identities are realities in the post-Internet fraud market. Relying solely on the traditional detection and investigation techniques is no different than the sales person who thinks their prospect hasn’t done any of their own research.
 Binary Nature of Fraud

Binary Nature of Fraud

Big data analytics and the predictive engines they spawned give Internet companies a way to monetize the online experience. By tracing online behavior, companies can target advertising, promotions, and point of sale opportunities based upon past buying decisions. Online habits of Web users can be associated with ideologies, interests, and values. Increasingly sophisticated probability engines predict future buying decisions with enough accuracy to fuel a dramatic increase in online sales and commerce over the past decade. Analysts use temporal versions of these tools to forecast market trends and evaluate risk.

In the fraud detection market however, early attempts to detect fraudulent behaviors using these same probabilistic engines have achieved limited success.

What makes detecting fraud different than detecting interests, values, and ideologies? The simple answer: Fraud is binary in nature–either a particular sequence of behaviors is fraud or it is not. For example, if an individual provider of medical services submits claims to an insurance company for 5,000 hours of services in a week (an instance from actual data), there had better be around 100 employees licensed to provide that service. If there are only three or four employees with the required licenses, the provider has committed fraud. Probabilistic engines struggle to detect fraud because they are not capable of modeling this“all or nothing”nature of violating a law.

At Pondera, we still make use of predictive analytics. But rather than detecting absolute fraud, we use the algorithms mostly to inform our fraud scores and to detect emerging fraud methods.

Once reliable methods of detecting fraud have been developed, predictive engines can also play an important part in helping insurance companies, financial institutions, and government agencies prioritize targets of investigation. Predictive models can identify the highest value targets that will recover the most money or disrupt the largest criminal organizations.
SaaS Procurement Recommendations

SaaS Procurement Recommendations

This week, my company is responding to an RFP for SaaS fraud detection services. While we are thankful for the opportunity to respond, the RFP and its process also illustrates the need for governments to adjust their procurement processes with the advent of cloud computing. After all, we responded to the RFI for this procurement over two years ago!

This means that the current solicitation is at least partly based on product capabilities from early 2014. While this might not be a big problem for traditional IT projects, this is a lifetime in SaaS. In fact, if a SaaS solution offered mostly similar functionality over a two-year period, I’d recommend not selecting that solution. Effective SaaS solutions push new features in days and weeks, not months or years.

With this background in mind, I’d like to propose that governments consider the following three modifications to their procurement policies. Some of these changes may require assistance from legislative bodies and funding organizations in addition to procurement professionals.

1. Reduce the time between RFI and RFP: This will help governments avoid building their requirements on functionality that has long since been replaced. SaaS functionality is a moving target – it’s supposed to be.

2. Smooth out funding over multiple years: Traditional IT projects required large upfront implementation costs followed by lower ongoing support, maintenance, and operations costs (assuming the initial implementation was successful). SaaS solutions spread the cost more evenly over time as the solution continues to improve.

3. Make sure your staff is ready when you award: True SaaS solutions can be implemented quickly, often in as few as 120 days. By the time you award a project, you should be ready to discuss security plans, access the required program data, assign staff (not just project staff but system users), and address many other details that could often be delayed in lengthy IT projects.
The Problem With Knowing What You Know

The Problem With Knowing What You Know

I bet you cna’t bvleiee taht you can uesdtannrd waht you are rdnaieg. Unisg the icndeblire pweor of the hmuan mnid, aocdcrnig to rseecrah at Cmabrigde Uinervtisy, it dseno't mttaer in waht oderr the lterets in a wrod are, the olny irpoamtnt tihng is taht the frsit and lsat ltteer be in the rhgit pclae. The rset can be a taotl mses and you can sitll raed it whoutit a pboerlm. Tihs is bucseae the huamn mnid deos not raed ervey ltteer by istlef, but the wrod as a wlohe.

The preceding paragraph, which has made its way around the Internet for years, can be really fun to share with friends. However, it also serves as a caution to anyone involved in fraud detection. In many ways, bad actors, knowingly or unknowingly, have depended on how the human mind works to perpetrate fraud schemes. Like the old expression goes, sometimes the best place for fraud to hide is in plain sight.

This is especially true in government programs that process massive amounts of transactions and must adhere to a staggering number of program regulations. Traditional “top down” systems can analyze large data sets and find nothing wrong (after all, the first and last letters are in the right place). “Bottom Up” systems, on the other hand, will identify individual problems (the word is scrambled) but may miss the patterns in the data (this entire paragraph is scrambled). A common example of this is the medical provider that always “flies just below the radar” by maximizing claim amounts and frequencies.

The best detection processes take both a “top down” and “bottom up” approach. They can identify individual transaction problems as well as identify patterns of bad behavior over time. In this way, you can make the old “80-20” rule work in your favor. 80% of improper payments are likely caused by 20% of program participants. If you only address each individual transaction, you’ll never run out of work but you also never really improve your program integrity efforts.

Click here for an infographic on the "80-20 rule".

Newsletter

Keep up on our always evolving product features and technology. Enter your e-mail and subscribe to our newsletter.

Email:

About Our Company

Pondera leverages advanced prediction algorithms and the power of cloud computing to combat fraud, waste, and abuse in government programs.



Get in touch

  • Sacramento Address: 11290 Pyrites Way, Suite 100, Gold River, CA 95670

  • Phone: (916) 389-7800

  • Email: info@ponderasolutions.com

Follow Us