Not long since my previous post, but a story published yesterday caught my eye upon which I feel compelled to provide comment. Most of my time over the past year has been spent speaking to regulators, investment banks, central banks and exchanges regarding
market manipulation in one form or another. Ultimately, the regulators and exchanges want to ensure a fair and open market place, the central banks want to provide a stable trading environment that won’t bankrupt the economy, and the investment banks want
to avoid costly fines and reputational damage. There is one glaringly obvious obstacle standing in the way of this at the moment – High-Frequency Trading (HFT).
Rather interestingly, the NASDAQ (the most recent victim of the problems caused by HFT) was the first electronic exchange to have been created, in the early 70’s. At this time, the rest of the world was blissfully writing documents on typewriters, calling
people using tethered home phones, and were exchanging messages using something called ‘the post’ (having researched this, it appears to be some form of paper-based system that requires human intervention).
Derivation aside, the way in which stock has been traded has gradually changed since the inception of the NASDAQ, and the speed-of-execution has increased. In itself, this was fine, because everyone was still playing on the same field, and ultimately to
execute a trade a human had to place it somewhere along the line. Then along came the wolf in sheep’s clothing known as Algorithmic Trading. At first sight, the ability to trade automatically without the need for interaction seemed like a great idea – in fact
I’m a BIG proponent of using analytics in this way and it yields optimal results in many different lines of business. However, allowing the computer to trade on your behalf also introduced a rather more dangerous variant of algo–trading, High-Frequency Trading.
With HFT, you are now able to process information and execute trades thousands of times a second; the kind of throughput a mere human could never aspire to. If each trade yields a fraction of profit, with enough of them you can start to build up significant
sums of money, hence the motivation behind using this technique. Additionally, the less scrupulous can use the technique to paint the tape to great effect. Place enough buy/sell orders in rapid succession outside the expected spread and the price of the stock
is going to shift as desired. Sadly, our exchanges are not designed to be used in this way – look at the Facebook IPO (still disputed), Knight Capital, and now Peet's Coffee and Tea. In each circumstance, HFT has been left unchecked and caused severe problems,
and in the case of Knight Capital, nearly bankruptcy. To summarise, the problems as I see it with HFT are as follows:
- Under heavy load, the exchange's software can’t deal with the volume/frequency of these HFT trades, and simply crashes
- The exchanges don’t currently have the software to detect errant behaviour either in real-time or near real-time to prevent circumstances such as the Peet’s Coffee and Tea share price soaring.
- HFT opens up a great opportunity, for those inclined, to manipulate and abuse the market
- If misconfigured, the algorithms can cause great damage to the investment company concerned, or even worse, the market as a whole.
So, what are our choices? There are two possible solutions I’m going to throw out there, both practicable:
- We cap frequency of orders & trades per trader to a pre-defined limit, of say 5 per second. Essentially HFT is outlawed, although this may be unpopular with the Exchanges given most trade activity (and hence their profit) is derived in this way.
- The exchanges invest in high-performance technology to monitor the markets in real-time and look for anomalies, known patterns and other triggers that indicate errant HFT activity on the exchange. This software exists, as it is exactly the kind of stuff
I deliver as rogue-trader surveillance, but is it the right approach to address the issue of HFT – I just don’t know.
Ultimately, most forms of Algorithmic Trading are valid and valuable evolutions in the trading arena, but HFT may simply be a step too far, incompatible with today's exchanges.