Teaching tomorrow's traders

September 16, 2016 12:00 PM
The trading universe has evolved from one focused on finance and “Type A” personality traders to a more scientific approach with quantitative techniques. The academic world is also changing to meet this need.

The University of Chicago’s Financial Mathematics program, FinMath, was founded in 1996 at the tail end of a period of rapid change in the world of trading that had been revolutionized by accessibility to ever faster, cheaper computing power.  

Wall Street was becoming increasingly electronic with the rapid transition from floor to computerized trading. As the dynamics of trading changed, firms were eager to add scientific knowledge to their trading desks. The University of Chicago, with a reputation as a leader in scientific research, wanted to help its newly minted PhD students in the hard sciences get exposure to quantitative skills that could be applied to the world of finance. With the number of traders with PhDs and trading firms run by PhDs still relatively few, the quest was on to provide Wall Street with all the Masters of the Universe they could handle.

The primary focus of the early trading curriculum was on providing a deep mathematical understanding of derivative pricing and the associated probability theory, along with a heavy dose of statistics. These topics allowed for the pricing of exotic derivatives, and applications toward pairs trading strategies. In addition, there were course offerings in fixed income mathematics that focused on yield curve modeling and quantifying fixed income portfolio risk and foreign exchange pricing, as well as the mechanics of forwards, repos and swaps.

Floor to screen

As the shift from trading on the floor of an exchange to computerized screen trading accelerated, the industry’s appetite for quants, particularly entry level employees, intensified. The trading firms were trying to take advantage of the higher computational capacity in addressing issues surrounding pricing functions in terms of accuracy and speed. Trading strategies focused on faster and more accurate pricing while identifying arbitrage opportunities and locking them in as quickly as possible.  

Today, the quest for greater speed continues to be a major trading industry driver.  As a result, FinMath’s curriculum has evolved to meet the dynamic needs of the industry. While the core of the curriculum remains, it has changed around the edges. The biggest difference between the demands of trading industry then and now revolves around the ability to code. The industry today demands traders who possess not only advanced knowledge of probabilities, statistics and mathematics, but also have solid programming skills. Quantitative knowledge is now only half of the equation for success. The other half is the ability to turn theory and ideas into working code. This requirement is of particular importance in high-frequency trading (HFT).

A need for speed

With its ability to allow traders to make money at microsecond intervals on price changes in fractions of a penny, HFT relies on speed and statistical analysis to drive returns. In order to provide entry level traders with the tools they need, the curriculum adapted to offer a market micro structure course that focuses on understanding the demands of algorithmic trading at extremely low latency. This includes volatility modeling using high-frequency data as well as stochastic programming techniques employed in high-frequency market making strategies.  

Because of the particular demands of HFT, the ideal entry level trader for an HFT firm must now not only have a solid command of mathematics and statistics, but he/she must also know how to make use of this via an ability to code the algorithms that will be conducting the actual trades.

Communications with hiring managers of Chicago based HFT firms yield a consistent message: They want applicants that not only understand the underlying probability and statistics that power HFT algorithms, but also understand how to effectively translate these concepts to working code. At the very least they (applicants) need to have command of basic scripting languages, such as MATLAB and R, where ideas can be generated and tested.

For trading firms operaing in longer time frames, topics such as Data Analytics and Machine Learning are in high demand.  Where once instinct and market knowledge drove trading decisions, now it is quantitative analysis that leads the way. The trading jocks from the floor have been replaced by the quants from math and computer class. Most trading today takes place in offices where quantitative analysts mine through mountains of data as quickly as possible to identify profitable trading opportunities.

The whole package

However, while the program fully embraces the move toward greater automation and quantitative analysis in trading, we understand that it is crucial for future traders to have a solid understanding of market logic. It is not enough to say the algorithm works because the numbers say it does. To fully grasp the risks associated with a trading strategy, one needs to understand the underlying logic of the relationships upon which it is built, why they hold and what could cause these relationships to break down.  

The numbers alone will not answer these questions, but rather an understanding of markets as mechanisms for price discovery driven by human players.  It is in this context that behavioral concepts are presented and ideas are proposed on how they can best be incorporated into trading algorithms.

This view is particularly important at traditional investment banks. Managing directors in trading at several bulge bracket investment banks in New York describe their ideal candidate as possessing, “solid understanding of how financial markets work along with excellent quantitative skills.” One added, “If you can add programming skills and an outgoing personality to the mix you have the perfect candidate.” 

In other words, they would like candidates that can bridge the separate disciplines, including the customer facing aspects, which are increasingly required to function effectively in today’s investment banking industry. 

Based on employer feedback, the curriculum helps alleviate a lot of time senior traders have to spend training new hires. Students are equipped with solid programming and statistical regression skills that allow them to hold their own in very theoretical discussions. Particularly, while working with others who have been in the industry for five years or more.

Changing student demos 

With the technological evolution of the trading industry, the types of students drawn to the FinMath program have changed. In the program’s earlier days, the students were typically existing PhD students or recent PhD graduates and working professionals. They were mostly comprised of Americans, Eastern Europeans and Russians. Today, most students are younger, and hail primarily from China as well as India. While some have advanced degrees, the vast majority come to the program directly from undergrad with little to no work experience.

Given most employers value some hands-on industry experience, it is critical FinMath students obtain the benefits of internships and research opportunities in the trading industry to fully enhance the benefits of the rigorous academic coursework. 

To address this, the program has instituted such things as Project Labs where students collaborate on projects provided by trading firms under the guidance and supervision of program professors.  
In response to constructive feedback from trading firms suggesting students were graduating with solid theoretical knowledge, but lacking sufficient hands on experience with trading, the program has helped students organize a Trading Club.  The Trading Club participates in several trading competitions offered by both industry and academic organizations. FinMath is also pursuing efforts to institute internal trading competitions that focus on operating in multiple time frames, markets and investment vehicles.

FinMath also has increased its focus on career development and today has dedicated resources that focus on helping FinMath students with obtaining these valuable work place projects, internships and ultimately securing trading jobs. On average, 20% of each graduating class will obtain a position with a proprietary trading firm as a quantitative or algorithmic developer or trader/trading assistant. These graduates will focus on mathematical modeling using coding languages such as C++, Python, MATLAB and R with applications toward option pricing, statistical risk management, portfolio management and portfolio asset allocation. 

In order to best ensure that FinMath’s curriculum remains cutting edge, we have established relationships within the trading community and among recruiters that provide us with an effective feedback mechanism, allowing us to continuously evaluate our course offerings and objectives. The goal is to quickly adjust to, and even anticipate, changes in industry needs to adapt the classroom experience and content to ensure our students are ready to perform at the very highest level. The Industry looks to us for the best talent and our goal is to deliver.

Looking further into the future, trading continues to evolve in terms of speed and processing power. Topics such Big Data, Artificial Intelligence, fuzzy logics will continue to surface as will more effective ways to model volatilities. As problems get solved, more complex problems arise. The amount of data that needs to be sorted and analyzed continues to grow and the tools to move through it quickly are being developed. The program has built an effective network, which allows FinMath to stay abreast of industry needs and, therefore, continue to mold the curriculum to effectively provide the trading community with the traders of tomorrow.

About the Author

Bernardo Jorge is Director of Corporate Relations for the Master of Science Program in Financial Mathematics & an Adjunct Professor in Algorithmic Trading at the Department of Mathematics at the University of Chicago.