Regulators both in the EU and the US have been tackling this definition and how to regulate (or not regulate) high-frequency trading. In the EU, regulation is more prescriptive, with the set goal of “curbing HFT.” The EU is using defined parameters to categorize firms as HFT or not.

In the US, however, there is even a chasm between the regulators and litigators. Litigators are taking a similar approach to HFT as EU regulators, focusing on HFT as an entity, which requires a strict definition. The SEC and CFTC, however, have recognized the difficulties in defining HFT firms and thus are using a more principle-based model to define HFT activity, understanding that defining HFT alone does not solve the issues in the electronic trading markets.

The point of this article is to identify the various approaches to defining HFT, recognizing similarities and differences among regulators and, more important, separating fact from opinion.

Regulatory Approaches

In the EU, two regulatory bodies have worked to define HFT: BaFin, Germany’s federal financial supervisor, and ESMA, the European financial regulator. The German HFT law passed in May 2013, with four main components affecting high-speed and algorithmic trading, including a definition of HFT with the intention that all HFT firms must register with the German authority.

Germany was the first to put in place a High-Frequency Trading Act, which sums up HFT in four parameters; firms need to meet all four criteriain order to be deemed an HFT:

1. Buys or sells for its own account as a direct or indirect participant (even non-member firms can qualify as HFT).
2. Uses an algorithmic trading strategy that requires no human intervention.
3. Uses infrastructure that minimizes latency.
4. Generates high intra-day message rates, which constitutes orders, quotes or cancellations.

BaFin further clarified the two latter parameters, because again, what may be one participant’s definition of low latency may not agree with another’s, and a similar argument exists for what constitutes high messaging traffic, leaving the rule subject to opinion. To avoid this uncertainty, latency-minimizing infrastructure was deemed a 10 GB transaction line in co-location and high messaging as averaging more than 75,000 messages a day over a 250-trading-day window.

Elsewhere in Europe, MiFID and MiFID II empower ESMA, the European market supervisor, to regulate financial markets in the EU, broadly covering both trading and investment firms and the venues where they execute. Under the most recent edition of MiFID II, published in June, the definition of high-frequency trading started to emerge as a subsection of algorithmic trading.

The parameters that are described in MiFID are similar to those in the German HFT Act:

• Low-latency infrastructure, such as co-location, proximity hosting or high-speed direct electronic access.
• System-determination of order initiation, generation, routing or execution without human intervention for individual trades or orders.
• High message intraday rates that constitute orders, quotes or cancellations.

Similar to BaFin’s move to further define parameters intended by the legislation passed in Germany, ESMA was tasked with further clarifying MiFID’s HFT definition and recently put forth two options to approach HFT:

Option No. 1 articulates clear parameters and looks a lot like the German definition of HFT:

• The firm’s servers where orders are initiated are in proximity, co-located to the exchange’s matching engine.
• The firm is using the highest possible bandwidth currently available in order to connect to the exchange; a bandwidth in the range of 10 GB/s would be considered among the fastest currently provided. However, ESMA notes the fact that a high bandwidth is subject to technological change and therefore should rather be covered in a qualitative manner.
• Firms that generate two messages per second over the entire trading day should be considered as using a machine/algorithm. On that basis, to determine the number of messages per trading day, it would be necessary to multiply the amount of seconds available per trading day (which may vary from market to market) by 2. The message volume should be determined on a rolling basis per trading day based on the previous 12-month period.

Under Option No. 2, each trading venue would determine a median daily lifetime of the orders modified or cancelled and then analyze which of its members/participants fall below the median daily lifetime of orders modified or cancelled for the entire market (which means that these members/participants are classified as HFT). “Daily” means that orders with a lifetime longer than one day should not be considered for these purposes.

The logic behind the second option is that the calculation already is conducted regularly at the trading venues and thus is perceived by ESMA as a method that cannot be easily circumvented. Additionally, it does not need to be revised frequently so as to keep pace with the latest technological developments, as is the case with the first definition (and the German definition). This calculation system has to be read in conjunction with the MiFID II provisions – that is, there has to be infrastructure to minimize latency (co-location, proximity hosting or high-speed DEA) and system determination of order initiation, generation, routing or execution.

In the US, the definitions are far less prescriptive and leverage more broad-based principles to define HFT.

[Related: “HFT – In Search of the Truth”]

The SEC’s definition of HFT, as found in its Concept Release in March 2014, is somewhat similar to the general outline of the German and EU regulation, giving the following general characteristics:

• Use of extraordinary high-speed and sophisticated programs for generating, routing and executing orders.
• Use of co-location services and individual data feeds offered by exchanges and others to minimize network and other latencies.
• Very short time frames for establishing and liquidating positions.
• Submission of numerous orders that are cancelled shortly after submission.
• Ending the trading day as close to a flat position as possible.

Unlike the EU regulators, though, which go into further details on the strict parameters that define HFT, the SEC does not further clarify the definition of HFT, nor does it give any indication that it will do so in the future. Instead, the SEC Concept Paper states: “Proprietary firms may engage in a variety of different strategies.” Rather than attempt any single precise definition of HFT, the Concept Paper focused on the particular strategies and tools that may be used by proprietary trading firms and inquired whether any of them raised concerns.

According to the CFTC’s definition of HFT, high-frequency trading is a form of automated trading that employs:

• Algorithms for each individual transaction without human direction.
• Low-latency technology that is designed to minimize response times, including proximity and co-location.
• High-speed connections to markets for order entry.
• High message rates (orders, quotes or cancellations).

Like the SEC, the CFTC seems reluctant to further refine the parameters of the HFT definition. As it stands, the definition is broad enough to encompass a big portion of market participants. However, the CFTC’s TAC believes that further narrowing the scope may lead to regulatory arbitrage and, more important, could not understand how a definition would be used for regulatory purposes. In lieu of defining HFT, the CFTC seems to be focusing on a principle-based approach for bolstering risk controls across market infrastructure as outlined in its concept release last year.

Finally, in the US legal arena, the DOJ is investigating HFT to see if it violates insider trading laws, the FBI is probing HFT firms for various trading activities, and the NY Attorney General is using the Martin Act to open up an investigation to see if HFT firms get certain advantages not given to other market participants. At this point in time, it is not clear what definition of HFT is being used by these offices.

While there are similarities between the EU and US on how they are defining “HFT,” there are slight differences with what they are doing with that definition. In order to define HFT, strict parameters need to be used, as seen in EU regulation, which will need to be adjusted over time as technology develops. Regulators in the EU have acknowledged this complication in the regulation. On the other hand, the US regulators – the SEC and CFTC – have acknowledged that defining HFT is a difficult task and to do so may not tackle the major issues brought to light in the market in the “Great HFT Debate.”

In the next article, we will look at the German High-Frequency Trading Act in more detail. Since Germany was the first regime to update regulation in the electronic trading era, it is important to understand how the rules were defined and their intentions.