Anthony Hilton: We really need people who understand tech

This incomprehension is important because flash crashes are happening with increasing regularity and across more markets. There was one in October 2014 in US Treasuries, another in March 2015 in the currency markets, another in August 2015 in the stock market, and again in October 2016 in the currency market. You might even argue that ethereum and bitcoin experienced a flash crash just last month — though that does not seem to have stalled their upward momentum. And to repeat Lo’s point, no one understands what is going on.

This has understandably caused some alarm among the regulators. Just last month the Financial Stability Board — the policy offshoot of the G20 group of the world’s wealthiest nations — said in a report that “applications of artificial intelligence and machine learning could result in new and unexpected forms of interconnectedness”. In plain English this means that it has noticed these flash crashes keep occurring.

It adds the “lack of interpretability or auditability of AI and machine learning methods could become a macro level risk.” This, I think, means they don’t understand why either, but it could be dangerous. This is particularly the case because many of these algorithms are based on data sets from recent years where there has been a historically low level of volatility. They could behave very oddly when things get choppy.

Nor do people fully appreciate the extent to which computers have replaced humans. Writing in the Financial Times last week Gillian Tett quoted a JPMorgan expert who said that only 10% of US equity trading was now originated by humans — the rest was from computer programs driving everything from exchange-traded funds to high-speed trading. Another source put the computer-driven slice a bit lower at between 50% and 70% for stock market trading, 60% for derivatives and 50% for US Treasuries.

Investment advice is also increasingly being delivered by computers rather than people. We constantly read about the development of self-driving cars and the need to make sure they are safe; we should be just as concerned about the safety of self-driving investment.

It can only get more fragile: Parker Fitzgerald, a consultancy which advises banks on how to cope with the digital revolution, said in a report published yesterday that these august organisations are pretty well doomed unless they rapidly transform themselves.

They may cling to the idea that they are a one-stop shop for financial services, but in reality their legacy of old technology and systems cannot be adapted to give today’s customers what they want and expect. The chances are that, as they struggle to adapt, the banks will lose more and more business to new entrants who operate largely outside traditional regulation, who can use the Cloud to house their expensive systems, blockchain to provide security and who are much better placed to mine customer data.

The implication of this is that rather than trying to fight on their traditional turf the banks would be better off re-inventing themselves as digital platforms. They could then provide a home for a range of third parties who will offer different financial products and services to the bank’s customer base.

Time will tell whether banks buy into the idea, but the point is that whether or not they do, that whole area of finance will also soon be as dependent on technology as the markets already are — and much though they might pretend otherwise, there are a limited number of people who understand how the technology works, and more importantly how it interacts through networks.

Parker Fitzgerald then said that all this technology is provided by fewer firms as weaker competitors fall by the wayside, and reminded us that the much-vaunted Cloud is also provided by an increasingly small number of firms — three account for two-thirds of the market and one, Amazon Web Services, has a 40% share on its own.

It is not meant as any reflection on these players to agree that the report could clearly be onto something when it concludes that the origins of the next financial crisis could easily lie somewhere deep in all that technology.