Research

Fraud Detection System to Protect Elderly People

How do you make sure your money is kept safe? This is a question that people will often ask themselves when they hear news that a new fraud scheme is doing the rounds. Many fraudsters are transparent, with their schemes lacking credibility and authorisation – but many more are highly skilled at stealing money from people’s bank accounts. It is not just telephone callers and email scammers that steal people’s money – debit and credit cards can be copied and numbers stolen too.

Elderly people can be at a higher risk of being targeted by fraudsters, with some international estimates suggesting that between 10 and 20 percent of elderly people will be targeted at some point. The result is huge financial losses for elderly people across the world – which also results in financial institutions losing significant amounts of money.

This is something that Cal Muckley, professor of Operational Risk in Banking and Finance at UCD Quinn School of Business, is trying to change. He is well aware of the issues that fraud presents to financial institutions and individuals across the world, but is acutely aware of the impact it has on elderly people. He notes that many elderly people are at a higher risk of being targeted by fraudsters because they may have greater wealth after a lifetime of earning, and are also more likely to experience cognitive impairment, such as dementia.

many elderly people are at a higher risk of being targeted by fraudsters because they may have greater wealth after a lifetime of earning, and are also more likely to experience cognitive impairment, such as dementia.

“It’s mind boggling,” Muckley says. “You’ve got people calling at the front door conjuring up work that has to be done in the house and then charging exorbitantly for it. You’ve got people coming in over the telephone professing to be with various companies or financial institutions who provide misleading advice. So, let’s imagine that one of these fraudulent individuals was successful. What you will have then is irregular, unusual patterns on the account of the elderly person, which may well breach the threshold and create an alert with their financial institution.”

The results of the study – and the team’s new machine learning algorithm – were published in the European Journal of Finance in December 2018, where they showed how they had successfully reduced the number of false positives in their financial institution. In the article, Muckley and his co- authors, Gaurav Kumar, Linh Pham and Darragh Ryan, note that this is the first systematic study of alert models designed to protect the accounts of elderly clients. This is despite the fact that elderly people are routinely targeted by fraudsters. In May 2018, Bloomberg reported that America’s elderly population is losing $37 billion a year due to financial fraud.

To devise their new approach, Muckley and his team used three different statistical models: logistic regression, random forest and support vector machine learning. Using these three models, they were able to come up with what they say is a better alert model that can help banks meet their ethical obligation to protect elderly clients. They estimate that 20 percent of elderly people in the US have been victims of financial fraud. Their study claims that the systems currently in place to detect fraud in elderly clients’ accounts are “inefficient” and can overlook “nuanced activity”. It is because of this, they claim, that financial institutions are flooded with huge numbers of false positives – fraud alerts that are not actually fraud at all.

But things are starting to change. There have been legislative moves across the world to help better protect elderly people from financial exploitation. In Ireland, safeguarding legislation to protect people from financial abuse is in the works in the form of the Adult Safeguarding Bill and the Vulnerable Persons Bill 2015. Meanwhile, in the US, the Bank Secrecy and Elder Justice Presentation and Prosecution Acts note that financial institutions have a duty to prevent the exploitation of their elderly clients.

Muckley argues that the benefits are manifold for financial institutions who implement a new model like theirs to detect fraud. Not only will it help to protect their elderly clients, it could also result in fewer losses for the institution, and could help to protect their reputation.

“Reputation is so key to these institutions given the importance of trust,” Muckley says. “The bank does have a duty of care to protect its vulnerable clients, including the elderly, and that is becoming increasingly clear.”

Muckley says that financial institutions have the data they need to better detect fraud in the accounts of elderly clients – but says that they could be using that data in “a much better way”. “We don’t have sight of course of precisely what other institutions are doing, but I will highlight that the institution we worked with is one of the largest financial institutions globally and is well recognised and has a very low fraudulent rate. I suspect that the kind of approach we are championing isn’t used much out there. That’s an awful shame, because these approaches can better protect the elderly.”

Professor Cal Muckley was in conversation with Patrick Kelleher (BA 2015, MA 2017), a freelance journalist