As members of one of the most heavily regulated industries, financial advisors undergo a significant amount of training regarding risk assessment and ethical advising. It wasn’t so long ago that trades were completed as a chain of conversations, with companies discussing their investments with their advisors and their advisors initiating individual trades. Many people still manually invested in their 401(k)s, and companies made sure to analyze their own stock portfolios with a regular basis.
But that’s all changed with the advent of self-driving finance: a practice as inevitable as self-driving cars — but potentially far more disastrous.
It Begins with Artificial Intelligence
At a certain point, signals weren’t enough for investors, who have realized time and time again that it is emotion that is most damaging to traders. In an effort to strip out that emotion, many large firms have instead begun to focus on computer-based trades.
Algorithms are designed to target specific trades, and then unleashed on the market. At any given time, up to 70 percent of the trading on equity markets is likely the result of a computer-generated trade.
In theory, this is ideal. Trading is completed based on set and controlled parameters which can be adjusted if the ROI of the trading goes down. Firms can compete to create the best algorithms, so competition within the market certainly isn’t dead — and the algorithms of course need to be fine-tuned and adjusted periodically.
Not only is there lower overhead because fewer people are involved, but immense and important trading can be done within milliseconds. This allows for investors to react to market timing that would be otherwise impossible to react to.
But it isn’t all good.
Devastating Downsides to Automated Trading
When it works, it works; but there are times when it doesn’t.
When major global events occur or unexpected market behavior, algorithms may be left in the dust. In the best case scenario, they may do nothing to capitalize on a unique circumstance. In the worst case scenario, they could do the exact opposite of what they should do, losing a firm an incredible amount of money.
But it isn’t just the firm that could stand to lose out. The other major issue with automated trading is that there could be bugs or glitches in the system. If there are, the system could initiate a slew of trades that go against not only the company’s best interests but the interests of the market.
Automated trading is, at its core, a trade off: it is able to engage in a multitude of preprogrammed responses very quickly, but it may not be able to react to anything that it does not anticipate.
Automated Trading in the Wild
Despite popular belief, the 2010 Flash Crash — the most major evidence for this case — was not caused by automated investing algorithms, but instead was a case of intentional fraud by an individual spoofing millions of dollars of contracts. Nevertheless, there have been a multitude of examples of failed automated trading.
In 2013, Goldman Sachs reported a loss of $100 million following a number of indications of interest sent as real trades. In 2012, Knight Capital lost $10 million a minute, amounting to a total of $481 million by the time the program was turned off.
These aren’t small trading companies, nor are they at-home investors purchasing “day trader” software; these are major companies that are making large scale mistakes. Each time these issues happen, they cause significant disruption to the market itself, generally in the direction that the trades are placed in. This also opens te potential for issues such as the 2010 Flash Crash, in which a single entity or individual can potentially cause widespread damage to the market by “tricking” other algorithms into responding.
Is automated trading as dangerous as analysts believe? It’s impossible to tell. But, as these algorithms become more complex and more organizations and individuals begin using them, the risk becomes greater.
Because the stock market is not a vacuum, each sequence of erroneous trades carries with it the powerful potential to crash the market. These issues cannot be anticipated. At the same time, increased regulations and education regarding the risks of these self-driving algorithms may be prompting companies to be more careful with their systems.
Regards,
Ethan Warrick
Editor
Wealth Authority
