What is considered as predatory algorithmic behaviour and how substantial are the risks for long-only traders?
All investors work to try and garner market sentiment and information before placing a trade, but predatory behaviour can be defined as that which purely seeks information without the intention to trade.
Post-MiFID, these risks have become more substantial, with fragmentation of liquidity and the availability of more granular market data being two major factors.
The advent of multiple trading venues has encouraged a new breed of tech-savvy high-frequency traders that may use the speed advantage they have to get an advanced look at the order flow of other participants on other trading venues.
This allows faster traders to capture the spread between the bid and the ask price by offering liquidity in small size to long-only institutions – a technique similar to that used by electronic market makers. They could also figure out the direction of large orders by piecing together post-trade data and use that information to build their own positions.
While such behaviour is typically associated with high-frequency traders, some market participants have observed that hedge funds can also employ momentum-based strategies.
The speed of trading venue infrastructure is also a factor. If data from one trading venue is disseminated at a slower pace than data from another that trades the same stock, this presents latency arbitrage opportunities.
But some market participants consider the debate around predatory algorithmic strategies to be overblown, pointing out that most algorithms today have inbuilt anti-gaming logic that gives them a degree of discretion to react to market conditions in real time.
What can buy-side traders do to mitigate these risks?
While their level of technological expertise is unlikely to match that of high-frequency trading (HFT) firms or hedge funds, there are a number of tools and techniques buy-side traders can use to limit the amount of information they give away when trying to trade in institutional size.
The first is simply to employ a diverse set of algorithmic strategies that would make it more difficult to figure out trading intentions. This should be done in conjunction with understanding the anti-gaming logic used in most broker-supplied algorithms.
Another is the use of real-time transaction cost analysis, a tool that allows traders to change the venue of execution or trading tactics for their orders based on fills they have just received.
In addition, dark pools, which were created specifically to help traders minimise the risk of information leakage, allow traders to post orders in an anonymous environment and can limit the ability of predatory strategies. However, there is a suggestion that some buy-side traders are naÃ¯ve about how safe dark pools actually are, as many encourage high-frequency flow.
What plans do regulators have to help protect long-term investors?
Both European and US regulators have realised that they need better oversight of automated trading, and in particular HFT firms, and a number of initiatives are in train to address this.
In Europe, HFT firms that trade using their own capital and exchange memberships are currently not captured by MiFID, leaving them subject to less oversight than other market participants. MiFID II seems likely to eliminate this loophole.
The next version of MiFID could also require all high-frequency traders to be classified under a new automated trading regime and may also obligate providers of algorithms to explain the design and function of their strategies to regulators.
From September this year US regulator the Securities and Exchange Commission will introduce a new large trader reporting system, which will require market participants that trade over two million shares or US$20 million during any calendar day, or 20 million shares or US$200 million during any calendar month, to report transaction information to regulators. This is likely to be followed by the introduction of a consolidated audit trail by the SEC before the end of this year, which will give the regulator access to real-time quote and order information in stocks and options.
Click here to vote in this month's poll.