When things go
wrong, the reflex reaction is to apportion blame. If we're late
to work, blame the train, the alarm clock, the kids - anything.
When things go wrong in the financial markets, there is no
shortage of bogeymen to hold to account. Call us over sensitive,
but at Automated Trader it does seem that the quant-trader has
joined the hedge fund manager, the private equity investor and
the good old-fashioned merchant banker (of rhyming-slang fame)
among the cast of pantomime villains paraded in front of the
public when the markets see red. Any sign of volatility and one
of the first reactions of the financial press is to pin the blame
on an algorithm or automated trading model for either causing the
problem or making it worse. (We look forward to hearing about the
first 'rogue trader' algorithm.)
Case unproven so far, but - as all model-builders know - if you don't filter out the noise, the future can become distorted. To put it another way, if you throw enough mud, some of it will stick. Should financial regulators, or worse, politicians, join in the blame game, the stakes will rise significantly. Will the first bankruptcy caused by losses traced to a 'malfunctioning' execution algorithm or trading model be swiftly followed by the kind of sledgehammer-to-crack-a-nut legislation (i.e. Sarbanes-Oxley Act 2002) introduced in the wake of WorldCom and Enron?
Canada's equity markets regulator, Market Regulation Services Inc. (RS), has already launched a shot across the bows of automated and algorithmic traders - and their brokers. In January, RS issued a statement that outlined the supervisory obligations of market participants "with respect to the use of an algorithmic trading system"1. As regulatory announcements go, RS's memo is more CYA than KYC. Indeed, RS admits explicitly its limitations to intervene to vary or cancel trades arising from a "malfunctioning" or "runaway" algorithmic trading system. (The term 'algorithmic trading system' is undefined by RS and as such may refer to orders placed by clients via bank-supplied execution algorithms or trades triggered by automated strategies.)
The regulator argues that because orders placed by brokers' clients via algorithms are effectively "entered on a marketplace without the involvement of staff" employed by the broker, they "present heightened risks" to the market and the participant, i.e. the broker. As such, RS calls on brokers to impose adequate "supervisory policies and procedures" to prevent regulatory violations by algorithmic trading systems. Moreover, the regulator suggests that brokers should ensure that algorithmic trading systems used by clients are "adequately tested assuming various market conditions" regardless of whether the system is "provided by a third-party service provider or by a client". Quite how brokers are to fund this latest increase to their compliance budget is not explained, but one message is clear: if your clients' algorithmic or automated trading activity causes a market disruption, we will blame you.
Of course, brokers already know this. As exchange members, they are already fully aware of the risks of high-frequency, computer-generated trading. Which is why they're working with exchanges to improve post- and pre-trade risk management tools and why those tools must be seen to work. If they don't, the risk of a regulator-imposed solution - such as market-wide risk management layer - increases considerably. And then brokers will have no one to blame but themselves.
1 Market Integrity Notice - Supervision of Algorithmic Trading, January 18, 2008 (Readers of a cynical nature might wish to substitute 'human' for 'algorithmic trading system' when reading this document.)