Governments in Canada could boost their revenues by one to three per cent by using data analytics to reduce fraud, waste and abuse, according to McKinsey research. For the federal government alone, one per cent in savings would amount to $3 billion.

It’s a tremendous opportunity for the public sector, especially now when budgets have been stretched to deal with the COVID crisis.  From tax collection to health care to emergency benefit payments, revenue losses can occur when fraudsters game the system or when process errors occur. At least half of the problems go undetected, says another McKinsey study.

“It’s so important because these programs impact people,” said Amanda Holden, Solutions Executive and National Practice Lead for Fraud & Security Intelligence at SAS Canada. “It’s critical to make sure that money is getting out quickly and into the hands of the people who need it.”

How can governments fix the problem?  

Game-changing advances in data analytics now give them the ability to monitor and detect issues before payments go awry. When it’s done well, McKinsey says it can pay a return on investment of ten to one, or more. Indeed, one European agency saved $344 million CAD in six weeks by applying analytics to its COVID paycheque replacement program, said Carl Hammersberg, Senior Manager, Government and Healthcare Risk and Fraud with SAS’ Global Security Intelligence Practice.

Governments must be proactive, said Omar Subhani, Director General and Deputy Chief Information Officer with Immigration, Refugees and Citizenship Canada (IRCC). “Our job is to get in front of the big challenges to be able to address them in a systematic way,” he said. “The use of analytics really helps us to identify those risks and get ahead of things.”

How data analytics helps prevent fraud

Banks use advanced analytics techniques to assess transactions every time someone uses their credit card or transacts online. These systems quickly bring together data on identity, location and spending patterns, for example, and assign a risk score as to whether the transaction is legitimate.

In the same way, public sector agencies can rely on data analytics as a first line of defence. The value is in connecting information to uncover unseen relationships in the data to identify potential risks. “By connecting data to create a network of information, analytics can find the risks and bad actors that the human eye normally can’t see,” said Holden.

A starting point in fraud detection is “bad list data,” advised Dan McKenzie, a Principal Solutions Architect in the Global Security Intelligence Practice at SAS. This list may include fictitious identities based on past experience or can be collected from a credit bureau or other external data sources. Since fraudsters are likely to vary their information when they return, the system will look at all of the elements, such as phone numbers or IP address, and flag any matches to the bad list. 

The Canada Revenue Agency (CRA) adopted a similar approach to find businesses that don’t file their tax returns. It uses data analytics tools to compare the information elements in web domain records to the existing taxpayer database to reveal the non-filers. “This solution has proven invaluable in performing the ‘heavy lifting’, thus freeing up time to focus on more meaningful pursuits in tax-risk analysis,”  wrote Jason Oliver, a CRA Senior Compliance Analyst in a white paper.  

Another powerful approach is to automate common controls and business practises and test them against the data.  A combination of several broken rules may mean a greater risk factor. Anomaly detection can also be used to spot “outliers” when comparing patterns among peer groups. For example, insurance companies use this technique to review billing by medical professionals. 

By using machine learning, advanced analytics models can become more effective over time. This happens when the system reviews its past results and automatically makes adjustments in real-time so that it can discover risks in a more predictive and preventative way.

Public sector organizations will see financial benefits by starting with analytics based on bad lists and business rules, said McKenzie, but “the holy grail would be to apply all of these techniques.” 

Improve service delivery and public trust

In addition to saving costs in the public sector, implementing data analytics can result in better and faster service to citizens. For example, it is used in child welfare programs to bring forward cases with the highest needs, said Holden. “It makes sure that overburdened social workers focus on the highest risk cases and get the right information quickly,” she said.

Similarly, Immigration, Refugees and Citizenship Canada uses analytics to manage volumes of applications to triage higher risk cases and to streamline the workload, said Subhani. For example, prior to the pandemic, IRCC was coping with a significant increase in temporary resident applications. As part of a pilot program, the system automatically reviews applications based on the eligibility rules and thousands of past decisions. It flags applications with risk factors for investigation by officers, providing all of the information they need in one place. The system adjusts to changes in the environment by feeding back information on non-compliant visitors. As a result, the time to process compliant applications has been reduced. It has improved productivity, while generating savings, according to a report on the project.

Ultimately, it’s about strengthening program integrity, said Subhani. “It helps us to identify the risks before things turn into front page news,” he said.

How to overcome challenges and get started

“While analytics is on the radar of many public sector leadership teams, many lack a clear roadmap for how to become more data-informed and advance the use of analytics within their organizations,” says Scott Sinclair, Deputy Minister of Crown Services for Manitoba in the Analytics Playbook.  

Like most large organizations, Governments face challenges in taking full advantage of analytics, from siloed data to privacy concerns. Data quality is a key success factor.  “You have to take care of the data,” said Subhani. “You need to spend an equal amount of time on data management and governance as developing the analytics engine.”

Privacy experts must be engaged throughout the process to ensure privacy by design.  “If you don’t bring them in early, privacy becomes a reason why you can’t do something,” Subhani said.

Organizations should start small, look for quick wins, and build from there. It’s important to develop a common vision on the project’s goals, said Subhani. “Leadership wants to see good results,” he said. “Don’t forget, you’re not just doing an analytics pilot. You are solving a business problem.” 

As the McKinsey study points out, leveraging analytics to solve business problems is a multi-million dollar revenue opportunity. For those that tackle the challenges, it says, “the benefits of improved government finances and citizen service delivery can be dramatic.”

 

Want to use analytics to save money in your organization?  Get more details on exactly how to do it, plus hear inspired success stories here.