The Economics of High-Frequency Trading: Taking Stock (2016)
Albert J. Menkveld
Annual Review of Financial Economics, Vol. 8, 1-24, URL/SSRN
This week’s AGNOSTIC Paper examines once again a somewhat controversial topic: high frequency trading (HFT).1I use this acronym interchangeably for both “high frequency trading” and “high frequency trader”. Full disclosure: I don‘t know much about HFT and the main motivation behind this post was to change that. So I did not choose the typical empirical paper but rather looked for an overview. The paper I found serves this purpose very well and is way too comprehensive for this post. Therefore, I will stick to the three points that are (in my newbie-opinion) the most important and interesting.
Everything that follows is only my summary of the original paper. So unless indicated otherwise, all tables and charts belong to the authors of the paper and I am just quoting them. The authors deserve full credit for creating this material, so please always cite the original source.
Setup and Idea
Before we start, let’s first look at what high frequency trading actually is? Unfortunately, there is not the typical high frequency trader and the industry got quite specialized over time. However, a common definition is something like “specialized trading firms that use high-performance computers and algorithms to execute transactions extremely fast”. They trade on very short time horizons (seconds, milliseconds, nanoseconds) and mostly engage in market making or “true” arbitrage transactions.2For example trading the same stock at different exchanges. Popular HFTs are for example Citadel Securities, VIRTU Financial, or Two Sigma Securities.
Given its need for computing power, HFT is a relatively recent phenomenon that mainly emerged over the last two decades. By 2014, the issue received a lot of attention when Micheal Lewis published “Flash Boys: A Wall Street Revolt”. In this book, he portrays HFTs not exactly positive and writes about a “rigged” stock market (more on that below). Even The Economist wrote in 2012 that it “[…] seldom finds itself on the side of restraining either technology or markets. But in this case [HFT] there is a doubt whether the returns justify the risk”.3The Economist is famous for its liberal and progress-oriented view, so this is really something special… The (public) image of HFTs is therefore mixed with a clear tendency towards negative.
On the other end, supporters argue that HFT reduces transaction costs, increases liquidity, and meets investors demand for continuous trading.4See for example this page from Deutsche Börse Group. But also keep in mind that Deutsche Börse profits from HFTs trading on their platforms… In addition to that, the boom of low-commission brokers like Robinhood & Co. would probably not have been possible without cheap trading via HFTs. So in contrast to their actions in early 2021, the reddit apes should actually like Citadel as it is the HFT market maker for Robinhood. But I am digressing…
Personally, I think a lot of the skepticism comes from the fact that HFTs are fairly secretive and extract a lot of profits for themselves. Therefore, I really like this week’s paper because the author sheds light on the issue in a more open-minded and scientific way. Admittedly, however, the paper is from 2016 and therefore not necessarily up to date. That said, I think there is a good chance that some of the more general results are still valid today.
Important Results and Takeaways
Trading costs strongly decreased between 2001 and 2011
Before he starts with HFT, the author first describes how the US stock market changed between 2001 and 2011 with respect to transaction costs and trading volumes. The following chart summarizes the result.
The key message is obvious. No matter which measure you look at, trading US equities became much cheaper during this 10 year period. Even more striking is the sheer magnitude of this cost reduction. Some of the transaction costs measures dropped by up to 80%. At the same time, the chart on the very right side shows that trading volume roughly doubled during that period. By and large, these results are positive because cheaper trading should lead to more efficient markets.5This is of course the welfare-maximizing view of economists. For conventional brokers who lost their commissions, the results are anything but positive…
None of these results directly indicate that HFT is good or that this cost reduction was only possible because of HFT. It is just an empirical fact. Nonetheless, the author highlights that those 10 years were a period of “[…] dramatic market structure changes for US equities […]” and that basically all trading shifted to electronic venues during that time. To throw away the scientific honesty for a moment and to give you a spoiler: HFT has of course contributed to this development and made a lot of it possible. But there are also some problems with it…
HFTs are fast, well-informed, and often market makers
Before looking at HFTs in more detail, it is important to first understand that most HFT strategies are essentially market making. This means that HFTs populate the limit order book with simultaneous buy- and sell-orders and earn the bid-ask-spread as a compensation for providing liquidity.6This may sound easy and simple, but implementing a HFT strategy in practice is of course insanely difficult. The author cites a lot of theoretical and empirical research that is consistent with this idea. So this is the first important insight to de-mystify HFTs. In most cases, they are “just” market makers.
To pursue their strategy, HFTs need both speed and information. In danger of stating the obvious, HFTs are indeed extremely fast. For example, the author quotes research showing that HFTs respond within microseconds.7Again, the paper is from 2016. Nowadays, HFTs may be even faster than that. Research also shows that HFTs seem to be very well-informed. For example, the author explains that HFTs not only submit a lot of orders but also cancel them when new information comes out. Another paper cited by the author finds that HFTs employ powerful text-algorithms and react to macro news within 0.2 seconds.
To sum up the results at this point: HFTs tend to be extremely fast, very well-informed, and mostly act as market makers. Nothing of this is inherently bad. In fact, having fast and informed markets who provide liquidity is actually beneficial for investors and certainly contributed to the strong reduction of trading costs. But as always in finance and economics, everything is a trade-off and some of those characteristics can also backfire…
Order-preying and arm’s races – it’s not all good
One concern about HFT is order-preying. The idea is quite simple. Suppose a non-HFT investor initiates a sizable trade that will have some market impact.8A large buy (sell) order will push the price up (down). Market impact means the trade becomes less favorable while executing it. If HFTs are able to predict such orders, they can use their speed advantage to update their quotes at the expense of the other trader (they “prey” on the order). In some sense, this is unfair because the other investor has no chance to react faster than the HFT and therefore incurs higher transaction costs. The author presents empirical research that is mostly consistent with this pattern. So order-preying seems to be indeed a problem but there are also ways for investors to protect themselves.9More about that on page 22 of the paper.
Another concern are “run games” and “arm’s races”. The idea is again simple. Suppose several HFTs coexist in fierce competition. If something happens in the market, for example a macro announcement, the HFTs will “run” to update their quotes as fast as possible. The fastest one will win and can extract some profits from the stale quotes of its competitors. The logical consequence is that HFTs continuously battle themselves with (costly!) technology to become even faster. If you apply standard economics to this situation, this is somewhat inefficient. The technological arm’s race is part of HFTs cost structure, so they will try to pass it to their customers. And since most HFTs are market makers, the customers are investors who will pay slightly higher bid-ask-spreads to compensate the HFTs for their technological expenses.
The author presents empirical research that is mostly consistent with such patterns. Needless to say, this is unsatisfying from a welfare perspective. Since most exchanges offer continuous trading, there is virtually no speed-limit for HFTs and the arm’s race continues forever. Although this clearly results from the characteristics of HFTs, the author also shows that exchanges could mitigate this problem by switching from continuous to some form of discrete- or batch trading.10Continuous trading means that a submitted order will be executed immediately. The opposite, batch trading, means that orders are lined up and will be executed at a specific point in time.
These are just two potential problems with HFTs and there are certainly some more. Nonetheless, I think these two are particularly interesting because both of them follow the idea that HFTs abuse their speed advantage at the expense of “normal” investors. While this is absolutely true and supported by the data, it is still no reason to demonize HFTs altogether…
The benefits seem to outweigh the costs
Once again, let me summarize the results so far. On the good side, we have HFTs as well-informed and extremely fast market makers. They offer liquidity and give investors the opportunity to continuously rebalance their portfolios which is clearly beneficial for markets. We also have seen that during the rise of electronic trading and HFTs, transaction costs across all measures dropped tremendously, sometimes by a factor of up to 80%. On the bad side, we have the problem of costly and wasteful “arm’s races” and unfair “order-preying”.
The author also discusses this pros and cons intensively and arrives at the following conclusion. If HFTs are only faster than others, they indeed hurt “normal” investors. This is Michael Lewis’ idea of the “rigged” stock market. The reasoning behind this is quite simple. If a HFT doesn’t know anything about the “right” price but observes your order, it may conclude that you know something. But since the HFT has the speed advantage, it can trade on the insight before you can. So your information moves into the price and becomes worthless for you. In my opinion, that’s indeed unfair and “rigged”.
On the other hand, it is not unfair if the HFT is also well- or maybe even better informed. For example, suppose the algorithms of HFTs correctly evaluate the short-term price effect of a corporate announcement. They will capitalize on this insight and use their speed advantage to update prices as fast as possible. This is helpful for other investors who may have not yet seen or evaluated the announcement as they will still receive a “fair” price in the market. So in this case, having HFTs as well-informed and fast market makers is clearly beneficial.
In reality, the positive effects seem to dominate. The author shows that calibrations of the negative “faster only”- and positive “well-informed market makers” effect suggest that the benefits of HFTs outweigh their costs. Admittedly, this was 2016 and the world has changed tremendously since then. However, since transaction costs have fallen again over the last years, I suspect that the results are still valid. Overall, as with most things in finance, there is no free lunch. You don’t get fast and informed market making without someone becoming rich from providing it and occasionally being unfair to a few people. But by and large, the paper suggests that we are better off with HFTs than without them.
- AgPa #76: ESG Myth Debunking
- AgPa #75: Optimal Investment Committees
- AgPa #74: Peer-Reviewed Research is Not Helpful to Predict Returns – Really?
- AgPa #73: Country and Industry Momentum
This content is for educational and informational purposes only and no substitute for professional or financial advice. The use of any information on this website is solely on your own risk and I do not take responsibility or liability for any damages that may occur. The views expressed on this website are solely my own and do not necessarily reflect the views of any organisation I am associated with. Income- or benefit-generating links are marked with a star (*). All content that is not my intellectual property is marked as such. If you own the intellectual property displayed on this website and do not agree with my use of it, please send me an e-mail and I will remedy the situation immediately. Please also read the Disclaimer.
Endnotes
| 1 | I use this acronym interchangeably for both “high frequency trading” and “high frequency trader”. |
|---|---|
| 2 | For example trading the same stock at different exchanges. |
| 3 | The Economist is famous for its liberal and progress-oriented view, so this is really something special… |
| 4 | See for example this page from Deutsche Börse Group. But also keep in mind that Deutsche Börse profits from HFTs trading on their platforms… |
| 5 | This is of course the welfare-maximizing view of economists. For conventional brokers who lost their commissions, the results are anything but positive… |
| 6 | This may sound easy and simple, but implementing a HFT strategy in practice is of course insanely difficult. |
| 7 | Again, the paper is from 2016. Nowadays, HFTs may be even faster than that. |
| 8 | A large buy (sell) order will push the price up (down). Market impact means the trade becomes less favorable while executing it. |
| 9 | More about that on page 22 of the paper. |
| 10 | Continuous trading means that a submitted order will be executed immediately. The opposite, batch trading, means that orders are lined up and will be executed at a specific point in time. |