The mysteries of bitcoin are turning heads yet again this week. At the start of the year, the digital token was worth $1,000; on Thursday, it breached $16,000. This is eye-popping, particularly given that so few investors actually understand how cryptocurrencies really work.
The mysteries of bitcoin are turning heads yet again this week. At the start of the year, the digital token was worth $1,000; on Thursday, it breached $16,000. This is eye-popping, particularly given that so few investors actually understand how cryptocurrencies really work.
But amid this frenzy, here is a more startling idea to ponder: what is happening with bitcoin is not actually the most head-spinning technological development in finance today. Far from it. Away from the public gaze, there are a host of other digital innovations emerging that have attracted little public attention, yet have more far-reaching implications.
Consider the structure of markets. A few decades ago, most investors assumed that markets were a place where human brokers traded securities, on behalf of flesh-and-blood investors, driven by strategies devised in their brains (or investment committees).
But today that idea is as quaint as assuming that currencies are controlled by a central bank. Marko Kolanovic, a JPMorgan analyst, estimates that a mere 10 per cent of US equity market trading is actually now conducted by discretionary human traders; the rest is driven by various rules-based automatic investment systems, ranging from exchange traded funds to computerised high-speed trading programs.
Of course humans write this code, and sometimes oversee trades. But at a recent financial technology conference at Michigan Law School, regulators and academics estimated that computers are now generating around 50-70 per cent of trading in equity markets, 60 per cent of futures and more than 50 per cent of treasuries. Increasingly, machine learning and artificial intelligence are being added to the mix, to analyse data, trade securities and offer investment advice.
What we are seeing, in other words, is the rise of self-driving investment vehicles, matching the auto world. But while the sight of driverless cars on the roads has sparked public debate and scrutiny, that has not occurred with self-driving finance.
This needs to change. Theoretically, digital finance could deliver huge benefits. As the Basel-based Financial Stability Board noted in a report last month, computers trade faster and more accurately than humans, and analyse bigger volumes of data to exploit price differentials. In good times, that should make markets more liquid and efficient.
But, as with self-driving cars, there is a catch: technology is moving faster than politicians (or voters) understand, and outstripping the legal and regulatory frameworks. Nobody yet knows how to assign liability if a self-learning financial program goes haywire. “How are we supposed to think about intent?” asks Yesha Yadav, a law professor at Vanderbilt University.
There are gaps in software laws. In the US, it is generally presumed that manufacturers have legal liability for product flaws. But as Washington’s Office of Financial Research has noted, “software developers are not generally subject to US product liability requirements”.
Another problem is regulatory fragmentation: although digital finance straddles geographical borders and asset classes, regulators do not. That creates a high risk that issues fall between the cracks. In turn, this fuels another issue: the technology is so fast-moving and opaque, that regulators find it hard to assess the cumulative impact or risks of contagion.
This is worrying. In recent years we have already seen some mysterious flash crashes, or sudden wild price swings, erupt in equity, bond, commodity and currency markets, apparently sparked by automated trading. This has not caused lasting damage, since these events were temporary and exchanges introduced measures to offset them in future. But nobody quite knows why these flash crashes keep occurring; and regulators admit that the arrival of AI will make it even harder to determine what is happening.
“Applications of AI and machine learning could result in new and unexpected forms of interconnectedness,” the FSB notes, adding that the “lack of interpretability or ‘auditability’ of AI and machine learning methods could become a macro-level risk”.
Digital evangelists will retort that since the arrival of the telegram, new technology has posed challenges for regulators; they also insist that the benefits of innovation more than offset the risks. Hopefully so. But the key point is this: just as we are scrutinising self-driving cars, we need to have a public debate about the computing revolution in finance. If the crazy antics of cryptocurrencies spur this, then bitcoin will have performed a public service.
