A report from the Economist Intelligence Unit

1Banks and big data: Risk and compliance executives weigh inHow big data can help banks manage riskHow big data can help banks manage riskA recent Economist Intelligence Unit survey of to avoid or reduce losses stemming from a variety bank risk management executives yielded a of risks—what counts is finding data that reveals surprising result: Over half of these senior retail, insights and leads to action.commercial and investment bankers say they lack Banks need to make the connection between sufficient data to support robust risk management. data acquisition, analysis and action. Insight is The assertion is especially noteworthy because always useful. But without action, it doesn’t lead to banks are awash in data, both big and small. The changes in behaviour. The survey confirms that growing complexity of transactions, volatility of bankers are having a difficult time with this critical markets, multiplication of devices, spread of aspect: Four in ten (38%) say that they face digitisation, multiplication of channels, demands difficulties in deriving actionable information from from regulators—all have increased the sheer existing risk data.volume of data exponentially. Yet there is not “Two problems come up again and again with enough to support robust risk management in retail bank data. First, it’s historical. And second, terms of improving risk management outcomes. it’s incomplete,” says Ozgur Kan, leader of Berkeley And when it comes to better outcomes—the ability Research Group’s credit analytics practice and Which of the following factors presents the biggest challenges to your organisation in improving its risk management outcomes? % of all respondentsLack of sufficient data to support robust risk management 51Difficulty in deriving actionable intelligence from existing risk data 38Inability to predict ROI from investments in new risk management techniques 37Difficulty of sharing data across silos within the organisation 20Lack of financial resources to implement new risk management techniques 14Concerns about regulatory constraints on integrating certain data sources 14Concerns about privacy of customer information 13Source: Economist Intelligence Unit survey, July, 2014.? The Economist Intelligence Unit Limited 2015

2Banks and big data: Risk and compliance executives weigh inHow big data can help banks manage riskformer head of GE Capital’s credit methodology and blocking fraudulent transactions almost function. “Let’s say you bought a house in 2011. instantly. The dimension that may bring the most You applied for a mortgage. You told the bank what value to the search for improved outcomes is you owned and what you owed, gave them your pay actually variety: the ability to support decisions stubs, and all the rest. Are you employed in 2014? with previously untapped forms of data.Did you sell your car to pay gambling debts? They “The variety piece is most interesting to me,” don’t know. ” says Mr Thomas. “We’ve always been able to store a The third-biggest challenge to improving risk lot of data. But now we have the opportunity to management outcomes is closely related to the really expand the variety of data that we draw on.” first two: 37% of bankers say they can’t predict For instance, Mr Thomas points to the ability to where the biggest return will occur when they need record, store and perform text analytics—statistical to decide where to invest in risk management. The techniques to model the information contained in world of analytics is clearly in flux, and the investor text—for every conversation between the bank and site AngelList.com catalogs several thousand its customers. In the past, many conversations analytics startups. Additionally, new analytics were recorded because regulators required it. But approaches pop up every day. It’s hard to know storage costs have now decreased so significantly where to devote resources when there appear to be there is no good reason not to record all customer so many viable alternatives.interactions. Not enough relevant data, difficulty moving If traditional transaction and counterparty from data to action, and the inability to know information comprises most of the data used now, which actions will yield the biggest return: As the unstructured data—text, audio, images or any data survey reflected, these are the three biggest that does not fit into a traditional database challenges to improving risk-management structure –represents a source of additional outcomes. But given this increasingly complex intelligence. According to a 2012 study by research landscape, what are the potential solutions?firm IDC, virtually all of the growth in data volume The fourth V is value comes from unstructured data, and only 0.5% of the world’s data is currently analysed. Unstructured Big data is often characterized by the so-called data relevant to bank risk management could be “three Vs” of volume, velocity and variety. But this images of fingerprints, legal text from mortgage terminology may obscure a fourth V—value. Some documents, or a raw feed of location data from a dimensions of data have more value than others. mobile device. All have the potential to provide new This value doesn’t come only from volume, as the insights, but all have been largely overlooked until survey participants recognized. Velocity is recently, as none come in a form ready to important in many situations, such as detecting incorporate into an analytics application.

? The Economist Intelligence Unit Limited 2015

3Banks and big data: Risk and compliance executives weigh inHow big data can help banks manage riskAbout the survey?In July 2014 the Economist Intelligence Unit management function and the rest working in the (EIU) carried out a global survey of 208 senior risk compliance function. and compliance executives at retail, investment Three in ten (29%) of the executives were from and commercial banks, with sponsorship from retail banks, about the same proportion (28%) SAP, seeking insights into how banks are using were from investment banks, and the remainder big data to improve risk-management and worked at commercial banks. Additionally, the compliance performance. Half worked in the executives were a diverse global group, divided C-suite, the others were at the VP or director level equally among North America, Asia-Pacific, with sixty-three percent from the risk-Europe, and the rest of the world. Benefits of a centralised approachwhole. Centralised analytics groups can contribute to this goal, but only to the extent that they Just over a third of the executives surveyed rated understand the business impact of risks throughout their banks as “well above average” in identifying the organisation and simultaneously build and analyzing risk. These banks are different from relationships with individual owners of business the broader survey population: Over half have lines within the bank. centralised analytics teams or centres of excellence “We are seeing a new generation of risk that develop best practices for their organisations, leaders,” says Justin Cerilli, a recruiter at Russell part of a broader trend in enterprise risk Reynolds Associates specialising in technology, management across all industries. The self-operations, and digital transformation at large identified high performers are moving towards financial institutions. “They aren’t necessarily centralised approaches.hands-on on the data side. And they aren’t Most bankers say that the single action that necessarily quants, as quantitative skills have been could most improve risk management outcomes is commoditised to some extent. Instead, they get “the creation of an enterprise-wide framework for the business impact of risk. And they can build stakeholders to achieve holistic perspective of all relationships across the organisation to get people risks confronting the organisation.” Managers with to take action about risk. Every conversation starts a holistic perspective understand how each part of with a discussion about risk, but it ends with a the business—and specifically the part that they’re discussion about leadership and change responsible for—contributes to the risk of the management.” ? The Economist Intelligence Unit Limited 2015