Big data: can it stop a crash? 

aat comment

Ian Fraser, author of Shredded: Inside RBS, finds out whether the clues to the next crash are hidden in the financial system’s data.

Big data brought us the Higgs boson, or ‘God particle’. It’s helping find cures for cancer, and it’s delivering autonomous vehicles. But can it improve auditing and financial services? Could it even prevent future financial crashes?

The vast majority of the data that swirls around most financial firms and accountancy practices still goes undocumented, effectively ending up on the financial equivalent of the cutting-room floor. But, thanks to dramatic improvements in computer power, algorithms, black-box thinking, data processing and storage, and blockchain technologies, it is now easier for firms to analyse, interpret, structure and extract value from data that was traditionally ignored.

Kara Stein, a commissioner at the US Securities and Exchange Commission (SEC), believes that big data may be the key to preventing future crashes. She told delegates at the 2016 Big Data in Finance Conference that the SEC has embarked on “an unprecedented data effort” that will culminate in the “largest data repository of securities trading activities in history”. She believes this data trove can be used to help ward off future crises.

Andrew Lo, professor at the MIT Sloan School of Management, suggests the financial system is becoming so complex, and evolving at such a pace, that new ways of regulating need to be found. For example, perhaps the focus should be on systemic risk rather than individual products and services.

Lo advocates an ‘adaptive market hypothesis’ that takes into account human behaviour, to supersede the increasingly discredited ‘efficient market hypothesis’, which posits markets are perfect, as well as “a new, interdisciplinary paradigm for modelling and predicting system-wide risk”.

Crisis averted?

Audit firms, which blotted their copybook by failing to spot the last crisis, could also play their part in foreseeing the next one by weaving big data into an overhauled audit process. Roshan Ramlukan, EY partner for global digital accounts and former global assurance analytics leader, says: “The audit of the future could look quite different from the audit of today.”

Writing on the firm’s website, he added: “Auditors will be able to use larger data sets and analytics to better understand the business, identify key risk areas and deliver enhanced quality and coverage while providing more business value.”

In finance, among the most advanced users of big data are hedge funds. One, London based Winton Capital, built its investment approach on delving into obscure data sets, such as wheat prices dating back 1,000 years, with a view to finding ‘buy signals’. “We’re data-hungry,” says chief operating officer Nick Saunders.

“Processing data at massive scale underpins everything we do.” Big data and artificial intelligence are seeping into the financial world on a smaller scale, too. Ultimately, they should bring improved insights and decision-making, which should, in turn, lead to better outcomes for consumers.

Fact or fiction?

However, sceptics worry that, because the real world of stocks and shares is complex and imperfect, trying to present it in a reductionist, synthetic way could be misguided. Atul Shah,
professor of accounting and finance at the University of Suffolk, warns that the lack of appropriate skills in the sector, as well as the amount of erroneous data out there, will undermine big data’s potential in the financial world: “There is a lot of fictional data and hype, which constantly distorts facts.”

Steve Keen, professor of economics at Kingston University and author of Can We Avoid Another Financial Crisis?, questions whether big data is a panacea or capable of delivering us from future financial peril. He does so primarily because individual banks and financial institutions lack access to rivals’ data, and because their choice of external data sets is subjective.

“What you get is an episodic data stream, not a systemic one,” he says. “It might tell you whether one particular lemming is running towards the cliff, but it wouldn’t tell you whether it’s already fallen off, or whether the herd of lemmings behind is advancing with such momentum that it will push the ones near the edge off the cliff anyway.”

This article appeared in our Jan/Feb 2018 issue of AT magazine.

Mark Rowland is a journalist and former editor of Accounting Technician and 20 magazine.

Related articles