Fears new algo rules will stifle innovation
There are fears that the introduction of measures to tighten monitoring of algorithmic trading in Australia will stifle innovation, particularly for strategies that rely on low-latency execution.
Just this week the Australian Securities & Investments Commission (Asic) published a consultation paper that, if enforced, would compel market players to step up risk control around algo trading.
They would be required to report annually that their systems had appropriate filters, and each time they made “material changes” – such as upgrading order processing or altering client order flow – they would need to confirm they had reviewed functionality. Failure to comply could result in a A$1 million fine.
Asic is seeking industry views by September 14, with the plan to publish new regulatory guidance on electronic trading the following month.
A pressing concern is that this deadline is too tight to allow the necessary investment in compliance. More significantly, it is feared that the need to provide annual notifications to Asic will have unintended consequences in terms of slowing industry innovation.
“If a client wants to execute on a new strategy that would require what the regulator considers ‘material change’ in our trading arrangements and controls, we would have to hold back to get these changes reviewed and certified with Asic,” notes a head of electronic execution at an investment bank. “This could cause the investor to hold back on their investment decision”
Asic would also require market participants to put automated filters in place. These pre-trade controls are meant to filter anomalous orders and manage extreme price movements at the level of executing brokers before they reach the two exchanges, ASX and Chi-X.
Asic has put the onus on market participants to suspend, limit or prohibit their algo-trading systems from putting through trading messages that could interfere with efficiency or integrity.
“We believe the benefits of the proposals are significant, with mandatory and direct controls helping to mitigate erroneous order entry and aberrant programs, and also helping to protect the market against the risk of automated trading systems exacerbating disruptive price movements,” Asic says in its draft consultation document.
But Alex Frino, a finance professor and CEO of Capital Markets Co-operative Research Centre, says that while the requirements on pre-trade filters might help to address the risk of HFT strategies exacerbating excessive volatility, it would have the unintended consequence of increasing order latency, inadvertently making HFT strategies unviable.
“There is no question that adding pre-trade filters would slow the market down,” he argues. “If you have to test whether the orders would cause undue price volatility before you release them to the market place, you would increase the latency of the order.
“This would make high-frequency trading algos unviable as HFT thrives on very low latency so that [traders] can quickly get into and amend orders in the market place.”
Frino suggests that low-latency execution plays an important role by providing liquidity, and ensuring prices in competing markets are not distorted.
He says the measures proposed by Asic would make all HFT strategies, both “good and bad”, less viable as trading would become more costly and risky for them.
“There is empirical evidence implying that shrinking the HFT community would have a negative impact on market liquidity,” Frino notes. “So the real cost of these new electronic trading requirements is on the entire marketplace.”