|
|
Although the investigation, undertaken during the autumn, acknowledges that high frequency trading can give rise to market abuse, it concludes that HFT per se does not damage the way the market operates.
The report will come as a relief to many trading firms which have been fighting a rearguard action against a rising tide of anti-HFT sentiment and growing regulatory scrutiny of the practice. The authors define a so-called black swan crash as an episode in which the share price deviates by more than 0.8 per cent over a timescale of 1.5 seconds or less - a timescale that exceeds the human capacity to intervene to override an algorithm, for example.
The researchers discovered some 18,520 examples of these extreme black swan episodes averaging more than one per trading day, so frequent as to suggest that black swan events may actually be nothing to worry about. But the researchers made another intriguing discovery: the pattern of black swan price jumps and slumps were generally found below the 650 millisecond threshold. In other words, the pattern of trading behaviour that takes place at speeds faster than 650 milliseconds does not appear to be replicated at slower speeds.
Full article (FN subscription required)