1.
몇 년동안 디지탈변환과 관련한 해외 사례중 Goldman Sachs를 주로 살폈습니다.
Goldman Sachs의 경쟁력과 IT현황
골드만 삭스의 IT 공개전략
Goldman Sachs, We build
그렇지만 데이타, 기계학습이 주제가 된 후 주로 다루었던 곳은 JP Morgan입니다. 그래서 JP Morgan이 어떻게 트레이딩부분을 변화하고 있는지 찾아보았습니다. JP Morgan의 변화를 보여주는 단어는 “Analytics, Automation & Optimization”입니다. 이런 업무와 관련한 직무가 Quantitative Research AA)입니다. JP Morgan의 구인홍보자료입니다.
The role of the QR Systematic Trading group is to identify opportunities to transform, automate and optimize our trading operations and to define and implement cutting-edge next generation analytics to support this business transformation. We cover all equities businesses and work closely with traders to develop data-driven solutions such as algorithmic strategies (high to low frequency), trading signals, risk models, portfolio optimization, flow categorization and clustering – and to combine them into automated trading processes or trading algorithms.
We are looking for a strong and experienced candidate to support primarily the APAC Delta 1 and Internal Market Making desks. Communication skills and drive are critical for this role as we expect the candidate to actively engage with the business and act as a culture carrier for modern data-driven methods and business automation.
Responsibilities
Build trading analytics and algorithmic trading strategies such as portfolio optimization, index arbitrage, statistical arbitrage and market making strategies (on Equity cash, Futures, ETFs) for the Delta One and Cash trading desks.
Identify business opportunities and contribute to the entire lifecycle from idea generation to production: perform research, design prototype, implement analytics and strategies, monitor daily usage and analyze performance
Support trading activity by investigating model and algorithm behavior (scenarios and post trade analysis, historical behavior)
Devise hedging and trading strategies and build execution logic
Quantitative Research AAO (Analytics, Automation & Optimization), Systematic Trading중에서
JPMorgan Arms Coders With Trading Licenses as Quants Advance을 통해본 위와 같은 변화의 결과는 다음과 같습니다.
“This is about convergence of the trader and quant”
“Looking forward, we will have much more automation.”
전통적인 트레이더와 퀀트의 구분은 없어지고 데이타와 소프트웨어 및 금융공학적 기반을 가진 전문가를 중심으로 트레이딩업무를 자동화하고 있음을 알 수 있습니다.
2.
몇 주전 JP Morgan의 Deep Hedging도 위와 같은 AAO의 연장선에 있습니다. 앞서 블룸버그 기사속에 등장한 사례도 Deep Hedging과 관계가 있습니다.
Sippel said his unit has also started developing machine learning-based tools for the trading floor giving the example of “RoboTrader,” a new tool to automate pricing and hedging of vanilla equity options.
지난 번 글에서 Deep Hedging: Hedging Derivatives Under Generic Market Frictions Using Reinforcement Learning을 소개하였습니다. 이 때 인용하였던 Risk.net기사를 보면 주식옵션에서 성과를 내었다고 합니다.
JP Morgan is using machine learning to automate the hedging of some equity options, a move that one quant calls a “game-changer”
무언가 작업이 있었다고 생각했는데 11월초 관련한 논문이 나왔습니다. Deep Hedging: Learning to Simulate Equity Option Markets 입니다.
We construct realistic equity option market simulators based on generative adversarial networks (GANs). We consider recurrent and temporal convolutional architectures, and assess the impact of state compression. Option market simulators are highly relevant because they allow us to extend the limited real-world data sets available for the training and evaluation of option trading strategies. We show that network-based generators outperform classical methods on a range of benchmark metrics, and adversarial training achieves the best performance. Our work demonstrates for the first time that GANs can be successfully applied to the task of generating multivariate financial time series.
이 논문에 등장하는 GAN와 관련한 모형은 저자인 Magnus Wiese이 발표한 Quant GANs: Deep Generation of Financial Time Series에 기초합니다.
혹 국내는 어떤 논문들이 있는지 확인해보니까 안양대학교 이우식 교수의 논문이 눈에 들어옵니다. 참고하세요.
