What is High Frequency Trading (HFT)?
You must have been hearing a lot about the High Frequency Trading these days, so thought I would share something on my blog to help my readers understand the basics of HFT.
High Frequency Trading has changed the trading landscape in last 10-15 years and more so after the onset of financial crisis. Ironically, HFT is very often blamed for the flash crash of 2007 and start of the crisis at the first place. But nothing has stopped HFT volumes to explode at the biggest exchanges of the world and ATS (Alternative Trading System).
People use the term HFT in different ways and for different purposes, which has made discussions about modern, electronic trading and its impact on markets more difficult. However, the term “HFT” is often used to refer to types of automated trading that “trade frequently.” HFT is also commonly characterised as having a sensitivity to speed (i.e., latency) and transaction costs; however, many types of trading fall within this broad characterization, including without limitation, modern electronic market making, short-term directional trading, and various forms of arbitrage and statistical arbitrage.
For most of these types of trading, low-latency, high-throughput trading technology is important. While these types of trading are often conducted on a proprietary basis, much of the algorithmic trading conducted on an agency basis uses the same technologies and tools as those conducted on a principal basis and for some purposes is referred to as HFT. Additionally, proprietary HFT is often, but not always, characterized as trading with relatively high rates of position turnover and small risk positions held outside of regular trading hours (relative to total trading volumes).
The growth of new electronic trading platforms and the increasing use of automation and advanced computing technology have raised questions about the effect these changes have had on the state, functioning and integrity of markets. Initiatives such as the U.S. SEC’s concept release on equity market structure in 2010 (U.S. SEC 2010b), the UK government’s ongoing Foresight Project on the Future of Computer Trading in Financial Markets (BIS 2011) and the European Commission’s MiFID II review in Europe (European Commission 2011), are all attempts to gain a better understanding of the true impact of these changes.
The following excerpt from a paper by Litzenberger et al. (2012) gives some basic ideas about the kind of strategies deployed in HFT:
Some HFT strategies trade primarily using resting orders, often quote two sided prices and rapidly adjust their quotes in response to market conditions. This style of trading is often referred to as “market making.” In theory, these strategies earn a gross profit from bid-ask spreads, which is partially offset by losses on their inventories due to adverse selection resulting from their quotes being traded with by those with better information (i.e., informed traders). In practice, bid-ask spreads (including any rebates or fees), adverse selection, other transaction costs and positioning are all important factors in the profitability of strategies relying on resting orders. Traders that use resting orders do not choose when they trade, but instead provide options to other market participants to trade with them. Adverse selection is mitigated by limiting the size quoted at the inside, quoting at multiple price levels and rapidly adjusting or removing price quotes in response to order flow, price, and volume changes. As such, messaging rates for strategies that use resting orders tend to be relatively high as these traders rapidly adjust their quotes to reflect changing market conditions and risk exposure.
Other HFT strategies, sometimes referred to as “directional strategies”, trade primarily with marketable orders, which differ from market orders because they have a price limit and are intended to interact immediately with resting orders. Directional strategies include mean reverting strategies that attempt to profit from transitory pricing errors and momentum strategies that attempt to profit from trends. Profits are generated when asset prices move favourably and sufficiently to exceed execution costs. The use of marketable orders implies a lower average messaging rate than trading with resting orders.
There is a lot more than this to HFT and I will try to explain as much as I can in my subsequent posts.
Need for speed to send the orders to the exchanges is changing the High Frequency Trading dynamics so frequently that only firms that can keep up with the constant and frequent technological updating of their infrastructure and trading strategies can survive in this game. Exchanges are gearing up for more recent radio-wave trading where orders would be sent through radio-waves which could be the fastest way to sending orders yet.
We are living in an age where milliseconds could be too long time and if you can’t keep up the pace then you are out of the race already.