What the DATA Act Means for Anti-Fraud Analytics


By Dave Williams, Former Inspector General, U.S. Postal Service

The U.S. government is now well underway on implementing the Digital Accountability and Transparency Act of 2014  (Public Law No. 113-101 official text), or DATA Act. The open data law mandates all 24 CFO Act federal agencies to first standardize, then publish, their spending information on one centralized accounting platform. The first reporting deadline is just months away, May 2017.

The DATA Act is not an end, it’s a means.

The DATA Act is arriving in the nick of time. The years ahead are unlikely to be a period of budgetary growth. The government pretty much has the resources they are going to have. The threats facing the U.S., however, are growing and coming at us at the speed of blur in an unforgiving environment.

The law promises an array of benefits for government management and public transparency. But none as beneficial and powerful as data analytics.

At first, agencies resisted the law, but most have come around and accepted it, and have even embraced it. Soon agencies will be saying that they had asked for it for many years.

However, agencies are now in a dangerous, and familiar, phase of play. It is reminiscent of another time when the government was, as usual, behind the curve and battling money laundering. Congress passed anti-money laundering legislation in the 1970s and 1980s, which required reports of suspicious transactions and codified compliance measures. The law required all financial institutions to file Cash Transaction Reports (CTRs) and international Currency or Monetary Instruments Reports (CMIRs) for large currency transactions. They dutifully began doing so, and a tsunami of paper converged on the Treasury.

For several years, the CTRs piled up and were boxed in warehouses. After a few years, Congress asked a simple question – “What are you doing with all those reports?”

It turns out that Treasury barely had enough time to box and store them, let alone run data analytics.  

An important reform that should have been the basis for great celebration, was instead an embarrassment for the government and led to resentment on the part of those financial institutions that filed the reports. Congress was not pleased.

So agencies learned their lesson and will never let that happen again, right? That, of course, is wrong.

Agencies may well be looking down the barrel of a look-a-like disaster with the DATA Act. Considerable resources are going into preparing the data definitions and schema.

Even still, it looks like we will mostly make it with a few usual suspects, like the Department of Defense, limping behind.

So, what’s the problem? Well, what agencies should have learned from the CTR disaster was that they need to begin to think in terms of how the government will use the new data.

A large chunk of the government’s resources should be put towards readiness programs so that the U.S. can take advantage of this new opportunity. Readiness levels and killer apps need to be in place the day the data reported. Actions need to occur immediately, not consecutively.

Every government agency, tech and private sector firm should be asking how will we use the new data. Once all agencies’ DATA Act data is reported to Treasury, we should imagine the government is done producing it. And the next question agencies should be asking themselves is what will we do with it?

Almost certainly, across the board, the answer is: engage in the big data analytics of the 21st century. The DATA Act will, for the first time, allow agencies to monitor their performance, detect and strike against fraud-sudden attacks.

With proper readiness levels and applications, the government will be enabled to cultivate a data-rich environment.

However, notoriously faulty instincts and intuitions may give way to evidence based decision-making and evaluation. But with the full implementation of the DATA Act there will be an array of government-wide benefits. The winners will be those agencies that fully utilize their data sets and work across government. Wasteful and duplicative programs will be cut, and cross your fingers, those involved in fraud.

Among the winners will also be the agencies conducting data mining. Bad data takes up roughly half the time of data analysts. The ETL (Extract, Transform and Load) process is mindless and soul crushing. That will be greatly simplified by the DATA Act. The agencies that are already embracing anti-fraud data analytics have experienced some pretty astounding feats already.

Stick around! The DATA Act is set to deliver the most valuable data set in the world. Clean and standardized data will increase current efforts by an order of magnitude. All those clever ideas for attacking fraud will suddenly begin to flourish. And if all agencies don’t join in, the participating agencies experiencing success will get smaller and smaller as they disappear over the horizon.

Data analytics pioneers have already previewed the road ahead. Even with terrible data, they are achieving impressive results.

Investigators at the U.S. Postal Service Inspector General opened 270 investigations in 2015 and recovered $20 million directly attributable to data analytics. The auditors using their models booked about $100 million in savings opportunities. Both the U.S. Postal Service audit and investigative data analytics programs are in their infancy and are bound to grow significantly under the DATA Act.

There will also clearly be a network effect as everyone joins in sharing their models and their findings. Intergovernmental initiatives can be enabled once the data is standardized and widely accessible.  

Our government is on the verge of something very big, but honestly if agencies are not going to get DATA Act-ready, agencies will receive some well-deserved criticism for going to all this trouble merely to stare at a new mountain of clean data. At some point others will stop asking for agencies cooperation and instead they will just need your desk.

DATA Act implementation hasn’t been smooth sailing, the government will encounter predictable hurdles that they should be working on now.

Washington is a secretive place, even when there are no secrets.

Huge departments are riddled with fraud and are in denial.

Procurement officials at the highest levels of the Pentagon believe or wish there is no fraud among the Defense contractors.

Program officials overseeing ballooning disability programs believe they are not being defrauded and both refuse to even look.

The agencies have a great deal of work to do in preparation for the first DATA Act reporting deadline in May. Data analytics could hold immense promise to ferret out fraud, find root causes, engage in real time continuous auditing, identify and manage risk, and increase the pace of agencies work. But only if Congress and the executive branch stick with it.