Four ways governments can stay one step ahead of fraud
Attempts to defraud the government have only accelerated over the past few years and governments need to stay adaptive and responsive if they wish to beat this escalating arms race, says Shaun Barry, Global Director – Risk Fraud & Compliance, at analytics software company, SAS.
Governments and fraudulent actors are engaged in an escalating arms race. Image: GovInsider
A recent report by global analytics company SAS found that fraud costs the UK government up to £53 billion (US$ 64.5 billion), amounting to around 6.4% of public money per year. And as government transactions become more digitalised, malicious actors are finding ever more innovative ways to exploit digital services and defraud governments in the process.
This is true in Asia as well. In Singapore, an organised syndicate laundered SGD $40 million (US$ 29.3 million) through government subsidies meant for skills training courses, reported The Straits Times. This was the largest fraud committed against the Singapore government to date and a key leader was sentenced to nearly 14 years’ jail this year.
Governments and fraudulent actors are engaged in an “escalating arms race”, says Shaun Barry, Global Director – Risk Fraud & Compliance at SAS. Here are four ways for governments to stay one step ahead of fraud.
1. A risk-based approach
First, Barry recommends adopting a risk-based approach to distinguish legitimate transactions from fraudulent ones made by malicious actors.
“The vast majority of transactions are legitimate, but there are malicious actors who have bad intentions,” he says. This includes organised crime outfits looking for weaknesses to exploit for a quick buck, and even foreign state actors seeking to destabilise other governments.
“Anytime you are giving money away is an opportunity for a fraudster to come in and say, ‘Hey, I'd like some of that money as well,’” he shares. This can include anything from social benefit programmes to business subsidies and carbon offset incentives.
Malicious actors can scour the dark web for private information to better impersonate citizens, or use generative AI like ChatGPT to create fake medical records to submit false claims, he shares.
This is why it is critical for government organisations to adopt innovative approaches to verifying identity and flagging out potentially high-risk transactions. Governments can learn from banks and financial institutions which have developed comprehensive approaches to identifying and monitoring different risks, he says.
Malaysia’s central bank, Bank Negara, is currently developing a National Fraud Portal that will use advanced analytics to curb online fraud, reported The Star.
2. Real-time analytics
Organisations can get started by adopting real time analytics that can immediately analyse the validity of transactions and flag out unusual behaviour.
As governments adopt digital transactions, they are expected to perform more transactions at faster rates.
“Now all of the electronic transactions are happening in sub-second response times. They’re very, very quick,” he explains. Analytics can help agencies quickly detect anomalous behaviour, such as if a citizen is suddenly submitting a transaction from a different country, an unusual device, or at an unusual time.
“One of the emerging approaches to fraud detection is to use that rich metadata that is captured but has often gone unused in the past,” he says.
When such a transaction is flagged, the algorithm can send that request up for review by a government official to determine if further action should be taken. This is similar to how banks and financial institutions are already checking the validity of transactions, he says. Singapore’s central bank is currently exploring the use of AI to identify suspicious transactions and stem money laundering, said chief Ravi Menon recently.
In Singapore, 11 people were charged for allegedly attempting to defraud government financial support schemes in 2022. Singapore’s tax authority, Inland Revenue Authority of Singapore, uses data analytics to review higher risk-cases of tax claims, and flags out suspicious activity to the police, reported CNA last year.
3. Cross-ministry data sharing
But such data analytics is only possible when ministries share data, he notes.
The government may be the largest collector of data in a country, but such data is often held in silos.
Agencies can consider “sharing that data strategically” to be able to identify bad actors and protect taxpayers’ money, he says.
Some governments have adopted privacy-enhancing technologies to better facilitate anonymised data sharing, reported GovInsider previously. In Malaysia, the government is driving data sharing efforts to improve citizen services, support policymaking, and drive healthcare efforts.
4. A responsive and adaptive posture
Finally, it is critical for governments to adopt a responsive and adaptive posture when it comes to the future of government fraud, as malicious actors will only keep developing more innovative tactics to defraud governments.
“There's continuing escalation and no resting on your laurels in this fraud detection business. Leaders need to give it the proper attention and proper investment to be able to keep pace with the fraudsters who are very innovative,” he says.
There is often a failure of imagination to consider how fraudsters might think, he notes.
It is also important to have safeguards in place, such as keeping a human in the loop, in case of an algorithmic failure. The Australian government recently gave a public apology for Robodebt, an automated debt assessment system that sent out thousands of incorrectly calculated debt notices to citizens from 2016 to 2020.
“Algorithms are most powerful not when they’re used to make a fully automated decision, but when they’re used by an individual to make better judgment calls,” notes Barry.
To learn more about how your agency can tackle fraud effectively, protect public services, and ensure taxpayers receive the value they deserve with an enterprise analytics strategy, download SAS' new whitepapers here.