High frequency trading is all of a go,
With joy to the traders and profits that grow,
It brings to investors an unhappy blow.
Investors and traders in capital markets have always sought to be better informed and to trade more quickly than their competitors. Thus new information technologies have been quickly adopted. As technology has evolved there has been a transition from carrier pigeons and semaphores, through telegraph, ticker tape and telephones, to fast computers and optic fibre links. The resulting evolution in capital markets has been towards faster and faster trading. High frequency trading, where milliseconds matter, is the current evolutionary outcome, but is this a good outcome for capital markets? Our evidence suggests not.[i]
The debate over the merits of high frequency trading is similar to the debate about the role of investors and speculators. Capital markets exist because of the needs of investors, but they also attract speculators. Speculators can be a good thing, they can add liquidity to the market making it easier and cheaper to trade, and they can move prices to their correct value more rapidly. However speculators can also destabilise markets. High frequency traders (HFTs) are analogous to speculators on steroids. Computerised systems are set up to trade with astonishing speed for quick profits. Such activity generates a lot of orders that don’t all turn into trades. HFTs are sometimes suggested to be preying on other market participants. However, they may also add to market liquidity, reduce the spread between the prices of orders to buy and sell (the bid ask spread), and may assist the movement of prices to their correct values. These are generally considered to be desirable outcomes for market quality.
One distinctive feature of HFTs is that the information advantage they seek is no longer fundamental information about the value of the company. HFTs use information in the pattern of orders and trades of other market participants to predict where prices are moving in the near future. Trading is a zero sum game and if HFTs are profiting, then someone else must be losing out. That someone is likely to be retail and institutional investors.
HFT strategies use the order flow entering the market to extract profits, whereas institutional investors use fundamental information about companies and economies to try and extract profits. HFTs profit from the order flow information arriving from the institutional investors. Essentially, HFTs are skimming cream off the top of an institutional investor’s fundamental analysis, but they win regardless of whether the institutional investor’s analysis was correct or not. It is therefore not surprising that some institutional investors have been crying foul and moving transactions off major security exchanges due to their concerns about HFT.
It is clear that HFTs profit at the expense of other market participants, implying a wealth transfer from investors to HFTs. However, if a new trader enters the market and is consistently profitable, they cannot instantly be condemned because they are more successful than their contemporaries. The more so if their trading leads to improvements in market quality. So what is the evidence on the impact of HFTs on market quality?
The evidence from carefully conducted research studies is mixed, but the weight of the evidence favours a positive contribution to market quality by HFTs. However, there are three concerns about this result. First, there is the difficulty of correctly identifying HFT trades and the risk of confounding the results of HFT with the results of other forms of computerised trading. Second, “improvements” in the metrics used to measure market quality may not really represent improvements in market quality. Third, the metrics themselves may be outdated. We will now consider these points in turn.
All automated computer trading systems are not HFT. For example automated trading systems may be used in executing large orders by breaking them up into smaller orders to minimise the impact on prices. In our review of the research literature we found that the positive findings on market quality often related to a combination of results from both HFT trading and other computerised trading.
Moving on to the market quality metrics. A key metric is the bid ask spread. A smaller bid-ask spread is generally considered to be a metric that signals an improvement in market quality making trading cheaper. However, there are trading strategies which will result in a tighter spread and which are also detrimental to the interest of investors. One example is the strategy of front running, or penny jumping large orders. The front runner submits a purchase order at a price one cent higher than the large order and so moves ahead of the large order in the queue for order execution. If the large order was previously at the front of the queue, the result is a tightening of the spread between bid and ask prices by one cent. While this strategy reduces the bid-ask spread it also makes it more difficult for other market participants to execute their trades.
The traditional metrics of market quality were derived twenty or more years ago, before the advent of extensive automated trading and when trading in open outcry pits was still a recent memory. As the distinguished researcher in security market trading, Professor Maureen O’Hara from Cornell, has pointed out markets are now fundamentally different. She has called for a new research agenda in the study of security market trading (market microstructure) and has talked of a new paradigm in response to high frequency trading. Seen in this light the traditional methods and metrics of market microstructure may need substantial revision and their contribution to measuring market quality may need to be re-evaluated.
In our analysis of HFT we focused on the impact on institutional investors and devised metrics to measure this impact. These metrics measure the expected time for an institutional order to be executed as a trade and also the probability of an institutional order posted at the best bid price, or best ask price, getting at least partially executed before it is cancelled or amended. Time to trade has become more important as the time interval has shrunk dramatically. Decades ago a millisecond was an insignificant amount of time in trading on an exchange, now it is considered a substantial interval of time. Who, even ten years ago, would have seriously considered spending $300 million to lay a high speed fibre optic cable, in order to cut the time it takes to send a message from New York to Chicago from 16 milliseconds to 13 milliseconds?
We used order book data for the Australian Securities Exchange (ASX) covering orders and trades over a five year period from 2009 to 2013, when HFT was growing strongly. We find that the probability of trade execution for institutional investors declined throughout this period, while the expected time for an order to execute continually increased. In another study we looked at what happened when the ASX offered a service, attractive to HFTS, which gave much faster access to order book information. After the introduction of this service, the probability of trade execution for non HFTs deteriorated, the costs of trading for retail and institutional investors increased and their exposure to predatory trading increased.
We have also looked at the provision of liquidity by HFTs. Do they provide liquidity on the side of the order book with plenty of orders, the thick side of the book, where liquidity provision is less needed, or on thin side of the book where there are fewer orders and a greater need for liquidity? Our evidence suggests that they more commonly supply liquidity on the thick side of the book and demand liquidity on the thin side of the order book.
Our conclusion is that high frequency trading is good for those that do it, but is detrimental to institutional investors and to retail investors as well. If the concern about market quality is concern about the interests of investors, then on balance HFT is bad for market quality.
[i] See the following two papers: “Is High Frequency Trading Beneficial to Market Quality?” by Richard Philip, Amy Kwan and Graham Partington, CIFR Paper No. 083/2015. Available at: http://ssrn.com/abstract=2673873 “High-frequency Trading and Execution Costs” by Amy Kwan and Richard Philip, working paper 2015. Available at: http://www.cifr.edu.au/assets/document/HFTExecutionCosts%20Kwan%20WP.pdf
The preceding post comes to us from Richard Philip, Lecturer of Finance at The University of Sydney Business School, Amy Kwan, Lecturer of Finance at The University of Sydney Business School, and Graham Partington, Associate Professor of Finance at The University of Sydney Business School. The post is based on their article, which is entitled “Is High Frequency Trading Beneficial to Market Quality?” and is available here. The authors would like to acknowledge support from the Centre for International Finance and Regulation (CIFR).