1.
고빈도매매.역사속으로 사라질까요? 사라진다는 판단을 하는 이유는 시장점유율의 하락입니다. 시장질서 교란이나 세금과 같은 외적인 요인도 있지만 내부적인 요인도 하락에 큰 몫을 차지합니다. 고빈도매매회사들의 경쟁입니다. 경쟁은 수익율의 저하를 가져오고 수익을 유지하기 위하여 투입하여야 하는 비용의 증가를 가져옵니다. 분명 고빈도매매를 단순히 High Frequency로 정의하면 사라질 수 있습니다. 그렇지만 나라별로 시장조작이나 세금 등으로 규제를 하기시작했기때문입니다. 그렇지만 알고리즘트레이딩이나 Low latency trading으로 바라본다면 현재 진행형입니다.
얼마전 블룸버그가 한국거래소의 마이크로웨이브정책을 놓고 보도한 기사입니다.
South Korea’s exchange is renewing its push to rid the country of a technology used by high-speed trading firms in the world’s biggest stock markets, arguing that it’s unfair. Korea Exchange asked brokers in July to disclose whether they’re using microwave networks, said Ko Young Tae, a KRX official in charge of derivatives policy. While Ko suspects the practice has stopped in the wake of an official bourse notice to avoid the technology last year, the exchange is surveying the market to know for certain, he said. Microwaves are the fastest way for traders to send orders to exchanges. While the technology has faced criticism around the world, the Korean exchange’s opposition is unusual for a national venue where the practice is legal. The bourse is focused on creating a more level playing field, despite high-speed trading proponents who say faster markets are more efficient.
“There is a possibility that the system violates fairness in markets,” Ko said in a phone interview.
High-Speed Traders Abandon Microwaves in Korea on Exchange Snub중에서
규제와 관련한 기사이지만 마이크로웨이브를 둘러싼 경쟁으로 읽힐 수 있습니다. 해외의 경우 마이크로웨이브뿐 아니라 스위칭장비를 둘러싼 경쟁도 치열합니다.
A handful of financial-technology startups are arming many of the world’s most powerful trading firms and exchanges with devices that promise to handle stock-market transactions at rates rivaling the speed of light, as the race for speed in financial markets remains alive.
Engineers at Sydney-based Metamako LP and Exablaze Pty. Ltd., and Chicago-based xCelor LLC are rolling out switches that take around four nanoseconds—four billionths of a second—for messages to transit from one side to the other, sending data from exchanges to electronic traders.
Trading Tech Accelerates Toward Speed of Light중에서
High-Frequency Trading Is Nearing the Ultimate Speed Limit에 따르면 국내에도 진출한 Metamako는 한달에 100여개의 스위치를 전세계에 공급하고 있다고 합니다. 한국에 진출한 외국인 투자자들도 애용합니다. 마이크로웨이브를 사용하고 있는 한국의 외국인투자자도 사용하고 있죠. 이들 제품이 제공하는 숫자가 85 나노초입니다. 85나노초? 그림으로 보면 이렇습니다.
Metamako’s hardware highlights the lengths that traders are willing to go to lower the “latency” of every component in their setup. The company figures it sells about 100 units a month at around $20,000 a pop—according to the Wall Street Journal’s piece on the company, that’s pretty affordable compared to many of the components in a high-speed trading rig.
2.
별 것 아닌 글이지만 글을 쓰기 위해 읽어야 하고 읽으면서 중요한 글들은 Pocket으로 관리합니다. 그동안 갈무리했던 글중 글머리의 주제와 관련한 고빈도매매를 다룬 논문만을 정리했습니다.
첫째 논문은 Expected Return in High Frequency Trading입니다. 논문은 1) opportunity, 2) capture, 3) effective spread, 4) effective rebate으로 이루어진 전략모형을 만든 후 수익율을 분석하였습니다. 이를 통하여 Latency가 수익율에 큰 영향을 끼치고 있음을 보여줍니다.
Defining α in high frequency trading is more complicated than in low frequency since not all strategies are based on price forecasts. More components are required, as is an understanding of the interactions between them. In this paper, we develop the α attribution model for high frequency trading by explicating its components and the trading tactics used to implement high frequency strategies. The results show why high frequency traders need to be fast in order to generate positive expected returns and why they are better at providing liquidity. We provide an example implementation using a sample of high frequency equity data.
두번째 논문은 How Rigged Are Stock Markets? Evidence from Microsecond Timestamps입니다. 미국자본시장구조와 관련한 논문입니다만 latency가 수익에서 중요한 의미를 차지한다는 사실을 보여줍니다. Best Execution을 위한 정보의 Latency때문입니다.
We use new timestamp data from the two Securities Information Processors (SIPs) to examine SIP reporting latencies for quote and trade reports. Reporting latencies average 1.13 milliseconds for quotes and 22.84 milliseconds for trades. Despite these latencies, liquidity-taking orders gain on average $0.0002 per share when priced at the SIP-reported national best bid or offer (NBBO) rather than the NBBO calculated using exchanges’ direct data feeds. Trading surrounding SIP-priced trades shows little evidence that fast traders initiate these liquidity-taking orders to pick-off stale quotes. These findings contradict claims that fast traders systematically exploit traders who transact at the SIP NBBO.
세번째 논문은 A high performance pair trading application입니다. 다른 논문과 달리 기술 논문입니다. 고빈도페어트레이딩시스템을 어떤 기술구조로 구축할 수 있는지를 다루고 있습니다. 아래 두 그름을 보면 논문의 성격을 이해하실 수 있습니다.(^^)
This paper describes a high-frequency pair trading strategy that exploits the power of MarketMiner, a high-performance analytics platform that enables a real-time, market-wide search for short-term correlation breakdowns across multiple markets and asset classes. The main theme of this paper is to discuss the computational requirements of model formulation and back-testing, and how a scalable solution built using a modular, MPI-based infrastructure can assist quantitative model and strategy developers by increasing the scale of their experiments or decreasing the time it takes to thoroughly test different parameters. We describe our work to date which is the design of a canonical pair trading algorithm, illustrating how fast and efficient backtesting can be performed using MarketMiner. Preliminary results are given based on a small set of stocks, parameter sets and correlation measures.
마지막 논문은 Does Pair Trading Work in Korean Market?입니다. 앞서 논문과 연결하여 한국시장에서의 가능성을 다룬 논문을 소개합니다. 혹 한국인저자로 한글논문이 있을지 살펴보니까 없네요. 그래서 영문을 그대로 올립니다.
We apply statistical arbitrage to conduct pair trading in the Korean stock market. We first construct a multifactor model in 5 selected sectors with the premiums from sector, size, value and momentum portfolios. Sector premium is the excess return of sector indexes over call rate. Second, we investigate whether the residuals from the multifactor model include predictable dynamics. Third, we implement pair trading considering the predictable dynamics of residuals and transaction costs. We control for standard risk factors and transaction costs, yet still find significant trading profit that prior literature cannot explain. Active asset managers can implement our pair trading strategies to enhance their portfolio performance. Our results suggest implications to both academic researchers and practitioners such as active fund managers, risk managers and traders.
