How much do you know about HFAT?
Disruptive trends like High Frequency Algorithmic Trading (HFAT) and machine learning are transforming financial services. And many firms, such as Goldman Sachs, are expanding their own algorithmic trading programmes as a result.
However, while HFAT offers numerous benefits (eg creating greater liquidity, increased volumes, reduced short-term volatility, narrower spreads, lower costs for investors, better price formation and faster execution of orders), it also carries considerable risks, which have the potential to cause widespread and significant market distortion, and create a disorderly market.
Algorithmic trading will face more scrutiny in the months ahead, with new obligations for all firms engaged in high-frequency trades under MiFID II. This isn't anything new - Germany's High-Frequency Trading Act was introduced in 2013 - but critics are sceptical about whether the measures will do anything to combat market abuse and question whether the time invested in data collection could be better spent.
Take a look at these five controls to help prepare for the new regime:
- Check the HFAT definition - there's never been a clear definition of HFAT until now and there's some dispute about whether the MiFID II definition is too broad. Check whether you fall under the definition of HFT - ie "trading in financial instruments where a computer algorithm automatically determines individual parameters of orders, such as whether to initiate the order, the timing, price or quantity of the order, or how to manage the order after its submission, with limited or no human intervention."
- Registration - firms engaging in HFT must be formally registered and meet the requirements of Regulatory Technical Standard 6 (RTS 6), demonstrating where software is purchased from, how it is developed, audited and stress-tested, how trades are monitored and the mechanisms for triggering alerts.
- Clarify terminology - get broad consensus across the firm about what constitutes an algorithm, algo trader and algo trading. For example, would you class a simulated stop order as algorithmic trading? This will ensure consistent application of the rules.
- Clock synchronisation (RTS 25) - is this technically possible? How can you overcome practical challenges? What management oversight is there over the time distribution chain? Are there systems in place to ensure a rapid response if problems are experienced? What checks can be carried out to ensure that everything is set up correctly? What sorts of events might cause time lags - eg maintenance or failed network switches - and how can you mitigate this? How can errors be detected? How can you ensure accuracy between data centres and servers running trading platforms?
- Check capability for record keeping - are you currently able to keep detailed time-sequenced records of all algorithms? What capability will you need in future?