The Future of Computer Trading in Financial Markets

Futurama’s Bender came to Earth with a dream: Kill all humans. Robots over the next decade may have a more pleasant goal: Make money for humans.

A natural corollary to this trend will be a sharp drop in the number of human traders, according to a working paper released by the United Kingdom’s Government Office for Science. “The Future of Computer Trading in Financial Markets” contained two key findings on threats to market stability.

First, self-reinforcing feedback loops could lead to considerable instability. “The effect of a small change looping back on itself and triggering a bigger change, which again loops back and so on within well-intentioned management and control processes can amplify internal risks and lead to undesired interactions and outcomes,” according to the paper.

Second, normalization of deviance could lull people into a false sense of security as “unexpected and risky events come to be seen as ever more normal” -- until catastrophe ensues, that is. The paper acknowledged that economic research has found no direct evidence of increased volatility caused by high-frequency trading.

Finextra has a link to the full report.

The Office for Science’s Foresight project generated the working paper. The project’s mandate included collecting science and evidence worldwide to objectively evaluate different perspectives on the risks and benefits of computer trading -- and predict how it might develop in the future.

“The use of automated optimization methods to design and improve autonomous adaptive trading algorithms is already commonplace in academic research, and its use in the finance industry looks set to increase over the next decade,” the paper stated.

“Because these next-generation trading algorithms will have little or no human involvement in their design and refinement, the behavior of any one such automated trader may be extremely difficult to understand or explain,” the paper continued. “And the dynamics of markets populated by such traders could be very difficult to predict or control.”

That doesn’t so bad -- as long as we can predict and control when they decide to kill all humans.