The financial services sector has always been a voracious producer and consumer of data. But in recent years, new sources of data, new techniques to analyse it, and new reasons to do so have combined to spur a wave of data innovation in the industry. And while concerns about privacy and bias may have discouraged institutions from pursuing that innovation, a post-pandemic push by regulators and clients towards responsible investment could provide an extra impetus.
The number of financially vulnerable adults in the UK has increased 15% since the pandemic. (Photo by Tolg Akman/AFP via Getty Images)
Data innovation in the financial services sector
One measure of the growing focus on data innovation in the finance sector is the growing frequency with which central bank executives mention the term “big data”. According to a recent study by the Bank of International Settlements, this has grown from a handful of mentions in 2015 to around 200 speeches every month that mention the phrase.
A survey of central banks by BIS found that 80% of central banks were formally discussing big data (meaning a combination of structured and unstructured, non-traditional data sets such as web content) and machine learning in 2020, compared to just 30% in 2015. They are using this data to understand economic trends, analyse monetary policy and measure financial stability.
Elsewhere in the sector, financial institutions are looking beyond the usual staples – personal financial records, company performance and asset prices – and using ‘alternative data’ such as satellite imagery, social media data and more. The most widespread applications of big data and machine learning in the sector are anti-money laundering and fraud detection, but institutions are using alternative data to gain new insights on investments and their customers.
This broadening use of data has accelerated during the pandemic. Not only have financial institutions accelerated the digitisation of their own services, the wholesale shift of consumer behaviour online has increased the volume of data available to analyse.
“Citizens have been forced to use more digital technologies, whether it’s for shopping, for Zoom, [or] more social media activity because of lockdown,” says Leanne Allen, a director in financial services technology and data at KPMG. “That has then increased the breadth of the amount of information available online, which means there are even more opportunities to leverage that data for financial services.”
Content from our partners
In some cases, Covid-19 revealed the shortcomings of institutions’ use of data. Many had to “throw all of their models out of the window,” says Matt Flenley, marketing and partnerships manager at data quality company Datactics, after overnight changes in behaviour invalidated the algorithms they use to detect fraud or calculate insurance premiums. One-way flights had previously been seen as a marker of fraud but became increasingly commonplace as national lockdowns were imposed. Insurance calculations based on the amount of time an average person spent in their home or the car were suddenly useless.
Data for the good of society
But perhaps the most significant shift resulting from the pandemic has been to the purpose behind financial institutions’ collection and analysis of data. “Over the last year or so, we’ve certainly seen a push from the regulators to actually use innovation and use data more for the good of society,” explains Allen.
Over the last year or so, we’ve certainly seen a push from the regulators to actually use innovation and use data more for the good of society.
Leann Allen, KPMG
After identifying that Covid-19 had caused a 15% increased in the number of financially vulnerable adults, the UK’s Financial Conduct Authority provided new guidance to lenders on how they should identify and serve vulnerable customers.
Data is likely to be vital for institutions in identifying at-risk clients. “Breaks in income, patterns of late-night spending, use of payday loan or betting services, or a change in spending behaviour may indicate the presence of vulnerability,” according to a report from real estate software vendor Altus Group. Other useful indicators include council tax payments and the use of pre-paid phone cards, according to credit ratings agency Experian.
Meanwhile, Covid-19 appears to have increased the need among investors to understand the social impact of their investments. In a survey by insurance provide Aviva, 55% of investors said the pandemic has made environmental, social and governance (ESG) factors more important to their decision making.
This will likely spur financial institutions to use a broader range of data to understand the ESG credentials of companies and other asset classes. This can include anything from analysing satellite imagery to track deforestation or social media listening to detect evidence of employee mistreatment.
Financial services data hesitancy
To date, however, the desire to be responsible corporate citizens has, if anything, discouraged financial institutions from pursuing data innovation aggressively. “The British banking system has such a strong, secure, safe reputation,” says Fedelma Good, who co-leads PwC’s Data Protection Strategy, Law and Compliance Services practice. “I think there was a concern that if you simply open the door [to AI and big data] then it would be difficult to keep that level of security and control and protection.”
These concerns have been exacerbated by growing algorithmic decision-making that has proven discriminatory. “We saw with the exam results fiasco last year, when all of sudden, all of the eggs are in the basket of ‘can a computer make a good decision?’ and if it’s not being done in a transparent explainable way, that’s just too much of a risk,” says Flenley.
Indeed, not all the discussion of big data by central banks has been positive, the BIS analysis of their speeches found. In 2015, almost all mentions were positive in tone; in 2020, 15% were negative.
But while concerns around privacy and bias are legitimate, data innovation can be a positive force in the financial sector, says Good. “Big data does not equal bad data,” she says. “We seem to have got to the point where it’s seen as Big Brother, and in fact, so much of it can be for the good.”
Growing demand among both regulators and clients for institutions to invest responsibly may therefore provide an extra impetus for data innovation.
