2016年-FCA英国金融行为监管局_occasional_paper_16_39页_2mb

FCA Occasional Paper 16 Summary

Core Content

This paper investigates the alleged predatory behavior of High Frequency Traders (HFTs) in the UK stock market. It examines whether HFTs can systematically anticipate the order flow of non-HFTs, particularly in the context of near-simultaneous orders and over longer timeframes (seconds to tens of seconds). The study uses a proprietary dataset containing detailed order-book data from 2013 for 120 stocks across three major UK trading venues: London Stock Exchange (LSE), BATS, and Chi-X, which account for approximately 85% of all on-exchange traded volume.

Main Questions

1. Can HFTs anticipate near-simultaneous orders sent by non-HFTs to different trading venues?
2. Can HFTs anticipate the order flow over longer timeframes (seconds or tens of seconds)?

Key Findings

• No systematic evidence of near-simultaneous order anticipation:
  The study does not find evidence that HFTs can anticipate and react to near-simultaneous orders sent by non-HFTs to different venues. This may be due to the physical proximity of UK trading venues and the regulatory environment that makes order routing less predictable compared to the US market.

• Evidence of order flow anticipation over longer timeframes:
  HFTs increase their activity before large non-HFT trades and when non-HFT buying and selling pressure increases. They tend to buy shares whose prices are expected to rise and sell those whose prices are expected to fall within the next 30 seconds.

• Uncertainty about the implications:
  The welfare implications of HFTs anticipating order flow are unclear. While such behavior could impact the prices at which non-HFTs trade, it is not straightforward to determine whether this behavior is beneficial or harmful to market efficiency.

• Limitations of the study:
  The results are specific to the HFT strategies analyzed and do not draw conclusions about other strategies. Additionally, the study does not investigate dark order books, which are not a major component of UK trading volumes.

Main Viewpoints

• Mixed perception of HFTs:
  HFTs are viewed both as liquidity providers and as potential market disruptors. Some argue that they enhance market efficiency, while others claim they exploit their speed advantage to gain unfair benefits.

• Regulatory and technological factors:
  The UK market's regulatory structure and physical proximity of trading venues may reduce the significance of HFTs' speed advantages. This is in contrast to the US, where HFTs are often associated with unfair market practices.

• Need for further research:
  The study cannot definitively conclude whether HFTs are anticipating order flow or simply reacting more quickly to public information. Additional research is required to distinguish between these two possibilities.

Key Information

• Data sources:
  The dataset includes detailed order-book data from 120 stocks across three UK trading venues. It captures all order submissions, amendments, cancellations, and executions at the millisecond level.

• Sample composition:
  The sample includes 60 stocks from the FTSE 100 and 60 from the FTSE 250, representing a wide range of liquidity levels.

• HFT definition:
  HFTs are defined as algorithmic traders that use proprietary capital and low-latency infrastructure. The study uses a business model-based approach to classify firms as HFTs, given the limitations of the data.

• Non-HFT classification:
  Pure non-HFTs are defined as firms that do not engage in low-latency proprietary trading or provide direct market access to HFTs. This group is considered more likely to be affected by HFTs' anticipatory behavior.

• Methodologies:

  • Near-simultaneous orders: The study analyzes HFT activity in response to non-HFT orders that arrive at different venues within milliseconds.
  • Longer timeframes: Two methodologies from academic literature (Brogaard, 2010 and Hirschey, 2013) are used to assess whether HFTs anticipate order flow over seconds to tens of seconds.

Conclusion

The paper finds no evidence that HFTs can anticipate near-simultaneous orders across venues, but there is evidence consistent with their ability to anticipate order flow over longer timeframes. The implications of this behavior for market fairness and efficiency remain unclear, and further research is needed to determine whether it is beneficial or harmful. The study highlights the complexity of assessing the impact of HFTs and the importance of considering the broader context of market dynamics.