Liming Brotcke is a quantitative manager of the Model Risk Oversight team at the Federal Reserve Bank of Chicago. She manages a team of model risk management specialists who conduct quantitative and qualitative review of a variety of models, and here discusses the fundamentals of what young quants will need to be successful.
The growing power and wide application of advanced modelling techniques, coupled with increasingly sophisticated algorithms and vast amounts of data, has made trading activities more complex and created a steady demand for quantitative analysts, otherwise known as quants. The increased speed of computers, dictated by Moore’s Law, has correspondingly increased staff productivity and competition intensity. The complexity of derivative trading and its impact on liquidity and financial market stability also draws lots of attention from regulators.
Traditionally, a quant is expected to develop and implement mathematical/financial models that evaluate derivative prices and trade securities. Depending upon their activities, quants can be further grouped into several categories. Front office quants work directly with traders and create pricing and trading models for traders. They design algorithms that search for alpha, the elusive returns above those returned as a component of standard stock market fluctuations. Front office quants have a greater emphasis on providing real-time solutions to specific problems than detailed modelling. Model validating quants take responsibilities of implementing and validating models from the front office to ensure error-free calculation used in trading. Research quants focus on creation of innovative pricing models and trading strategies for the front office. Asset management banks employ investment quants that leverage either exclusively quantitative strategies or a combination of quantitative and fundamental methods.
Understanding the underlying problem is critical for a quant to succeed, and requires a good grasp of how financial market works as a whole.
There are also quant developers who are responsible for programming, script writing, code debugging and other technical issues. Statistical arbitrage quants working at hedge fund firms research the inefficiencies present in the market with the help of automatic trading algorithm. Finally, capital quants are tasked to forecast banks’ credit exposure. In the United States, quants who don’t work directly with traders are also referred as “middle office” or “back office” quants and typically work in risk management.
There are a variety of activities quants often engage, including bonds, commodities, credit, equity, foreign exchange, and hybrids. Business needs also differ among commercial banks, investment banks, hedge funds, technology service companies, and software issuers. Roughly speaking, the buy side (i.e. hedge funds and alike) is more statistics and algorithmic trading oriented whereas the sell side (i.e. investment banks and alike) is more like traditional derivatives pricing and all that follows.
Quant activities will encompass solid understanding of business needs, the underlying theory (mathematics, statistics, data science, etc.), available or developing technology, and the regulatory environment.
First, a solid mathematical background is a must. The ability to work with stochastic partial differential equations, linear algebra, stochastic calculus, and probability theory to solve problems and derive solutions is the most desirable skill set to have. As such, a doctoral degree in highly numerate fields such as mathematics, physics, engineering, economics with a focus on econometrics and time series, computational finance, and so on is preferred by most firms. In the last decade graduates with Master's degrees in mathematical finance, financial engineering, operations research, computational statistics, machine learning, and financial analysis gain sufficient popularity among employers as well. With the wider adoption of data science and machine learning in portfolio risk modelling and portfolio management there is also an increased demand for predictive analytics in the recent years.
Second, quants are expected to command adequate financial market knowledge. The supervisory view of model is that models are simplified representations of real-world relationships among observed characteristics, values, and events. Understanding the underlying problem is critical for a quant to succeed which requires a good grasp of how financial market works as a whole, as well as sufficient familiarity with individual products being involved such as commodities, interest rate, foreign exchange, and asset-backed securities, aside from robust mathematical and analytical skills. Unlike the mathematical knowledge that can be learned in an academic setting, most quants acquire financial market knowledge by working with traders, portfolio managers, and other colleagues in risk management.
Third, software skills are also critical for quants to perform at a proficient level. C++ is the most important programming language and commonly used for high-frequency trading applications. Other statistical softwares include Matlab, SAS, S-PLUS/R, SQL, and Python. Advanced skills in Excel are also required, in particular when there is an integration of Java, .NET and VBA with Excel.
It is important for quants to add knowledge of regulatory requirements to their tool kit in the changing environment.
The most recent financial crisis has resulted in heightened attention on models developed and used for the financial market from regulatory agencies. The issuance of the BOG-FRB SR 11-7/OCC 2011-12 establishes supervisory expectations on key elements of model risk management including model development, implementation, and use, model validation, and model governance, policies and controls. More quants employed by banks have been involved in the control function often called model risk management or model validation unit, which validates the entire process of model development and implementation, approves or rejects the models before the front office can deploy them for production use. There are also specific expectations on financial institutions that leverage vendor models. The impact of such regulatory guideline is profound. For example, models cannot be used for business prior to validation and approval. For models already used in in production but with identified major deficiencies timely remediation plans and exception approval need to be in place. It is important for quants to add knowledge of regulatory requirements to their tool kit in the changing environment.
Maintaining large interconnected ecosystems of in-house and vendor built models will be necessary for banks as a going concern to meet regulatory requirements; correspondingly, the skill requirements will be ever diverse relying on people of varied educational and professional backgrounds. Lastly, developing communication skills will become more important in the future. Successful quants will have to communicate abstract mathematical constructs to end-users, senior management, business heads, and regulators.
Are you a quant looking to connect with the larger derivatives community? Join 150+ academics, financiers, regulators and technologists at Global Derivatives USA >>