Risks and Misuse of Big Data – Learning from Bank Fraud

Paul Gibbons big data

In 1995, I was tapped on the shoulder at Coopers (now PwC) by a partner in the forensic accounting practice who was leading our investigation into a 93 million pound fraud ($140m) at NatWest bank. “Gibbons, you traded options, did you not?” “Some”, I replied. Then, suspiciously, “It says here you speak French?” Once he had confirmed all this, the other shoe dropped. “We need you as an expert to interview this Frog (French person, derogatory) who seems to have lost Nat West a few quid (English understatement, ‘a fortune’).”

48 hours later I was to be an expert investigator/ examiner on a highly technical subject. As my finer days trading options were some 7 years prior, I quickly ducked into a bookstore and spent the next 47 hours reading Hull’s mathematical tome on options theory. While I had traded options (for years), my interviewing a professional trader was like a physics graduate student trying to understand whether Stephen Hawking had pulled the wool over our eyes with Black Body Radiation. Nevertheless.

The meeting included the CEO of Nat West (Sir Derek Wanless), the head of risk management, several executive committee members and our Coopers partner. What horrified me most about the meeting was that apart from the trader and me, not a single participant at the meeting seemed to understand the first thing about options (which are very intricate mathematically). My knowledge was rusty and rudimentary compared to that of a true expert, yet compared to the rest of them, I was a savant.

National Westminster (one of the largest banks in the world) was managed by senior leadership who were blind to how granny’s savings, and the country’s pension funds were being invested. “Under the hood” of this world-class bank was a very technical (mathematically) risky enterprise (options and derivatives trading). They had no more idea how to lead that than they did how to race a Ferrari. Senior leaders of businesses that become technically complex (as banking did in the 1990s and 2000s) cannot be complacent and assume that knowledge sufficient to prudently approve an international loan quickly translates into understanding the risk of derivatives. (One is trivial, the other complex.)

Could Big Data be like that? Again, we have a very complex technical subject with much opportunity to spend eight figure sums that will be wasted without the right governance, the right analytics culture, the right leadership skills, and the right mindset (data-driven, evidence-based). Ask Big Data the wrong questions, and you will get useless answers. Without the right kind of hacker mentality to explore unknown unknowns, corporations risk purchasing expensive systems that do not provide adequate returns. Given my earlier experience, and my recent toe-dipping into the world of machine learning, Hadoop, and R, there could be a world of pain if business leaders are unable to create an organization and culture that can translate the esoteric mathematics into trustworthy business insights.

Leave a Reply

Your email address will not be published. Required fields are marked *

Connect With Paul

Join Paul's newsletter for the latest provocations on philosophy in business.

Sign Up

Get the latest provocations in science philosphy and business from Paul.

We respect your email privacy


Connect with Paul

On Twitter

On Youtube

On Facebook



By using this form you agree with the storage and handling of your data by this website

Download Media Kit

View Media Kit