Co-founder of @AuroraSignals, quant hedge fund co-manager, financial advisor, entrepreneur, trader of equities | forex | options

Wed 10 September 2014


Twitter Logo Facebook Logo Google Plus Logo LinkedIn Logo Mail Logo

Goodbye Wall Street bankers, hello scientists.

Below is an actual Wall Street job posting recently received by email for a position as a Quantitative Analyst:

Requirements: excellent financial modeling and exceptional academic background in scientific field. PhD/ MS degree in statistics, applied mathematics, computer science, econometrics, data science, engineering or hard core science from a top university. 3+ years of directly relevant industry experience with equities.

  • Compensation: $300,000
  • Min Education: MA/MS
  • Min Experience: 3 years

In the 80s and 90s most Wall Street traders would have read this posting and thought, “why in the world does a firm want someone with a scientifically-based degree?” The answer is simple, the financial system works much like the natural world in many ways. Countless mathematical techniques used in day to day processes have been found to be exceptionally applicable to problems traders on Wall Street have been trying to address for decades. Now, in the era of computers, all of this is possible with this new breed of traders.

This new breed isn’t as new as you would think, however. In fact, computers have been aiding in trade decisions and order executions for about four decades now. It wasn’t until recent years that the media began covering the true evolution of Wall Street, from what it was in the 80s and 90s to what it is today. Gone are the days of a packed New York Stock Exchange trading floor with men in suits yelling back and forth to “buy” and “sell,” frantically waving order tickets in their hands. Over the last ten years, many of these human traders have been replaced by tens of thousands of servers, many of which reside in massive warehouses across the Hudson River in New Jersey, not New York City .

Times are changing, are your investments keeping up?

What algorithms do and how they’re used:

Decoding Wall St

Source: Decoding Wall St — Infographic / QuantConnect

Algorithms run the world around you

Whether you realize it or not, you used algorithms today to complete countless tasks. What would your average day look like without them? Most likely frustrating and even more time consuming than it does now. Drive a car made in the last two decades and you’ve employed the work of hundreds, if not thousands, of algorithms that do anything from monitoring your current position (GPS) to providing you with traction in rainy conditions (ABS, traction-control, etc.).

Do you own a smartphone? The “autocorrect” text feature is just one of many algorithms used to better your user experience and cut down on time to complete certain tasks. Did I mention algorithms also help keep us alive, or avoid potentially disastrous situations? Airbag systems are one such example. In the medical field they are used for test selection, diagnosis, therapy and prognosis, and automatic control of medical equipment, just to name a few applications. Source: Medical Algorithms

The image below displays the results of an algorithmic program medical scientists used to discover thousands of new side effects and problematic drug interactions. Wild looking, right?


Source: CREDIT: Image courtesy of Science/AAAS

On average, the program found 329 side effects for every drug, when most drug labels list an average of 69 effects. The new program, developed at Stanford University, combs through millions of doctor and patient reports to the U.S. Food and Drug Administration, Canada's MedEffect and similar databases.” Source: Tech News Daily

Financial algorithms and their performance

Broadly speaking, this same approach is being used by top Wall Street firms to effectively study patterns among stocks and to execute millions of trade orders every day. In fact, the use of algorithms in financial markets has become so popular that approximately 75% of all trades are now performed by algorithms. Collectively, the top wall street firms manage trillions of dollars through the use of algorithms. There’s a reason the brightest and most successful money managers in the world use algorithms to discover investment opportunities and execute on them. Look below to see how effective these programs can be at generating investment returns compared to buy-and-hold index investing.

Decoding Wall St

Source: Decoding Wall St — Infographic / QuantConnect

The algorithmic advantage

Why are algorithms becoming so popular among top Wall Street firms? When it comes to speed of thought, data processing, and zero emotional hindrance, the advantages become apparent. There’s a reason the top algorithmic investment firms have been able to make returns that dwarf those of their conventional, less successful peers. Arguments can still be made for human-driven strategies based on multi-decade-old approaches, however, the proof is in the numbers, and numbers do not lie. The brightest minds in the world are behind this movement and it’s important that retail investors understand the implications and adapt. Below is an illustration that lists some of the reasons algorithms are now dominating in the investment world:

Decoding Wall St Decoding Wall St

Source: Decoding Wall St — Infographic / QuantConnect

Types of financial algorithms

While the performance advantages are something to note, one of the most powerful advantages of algorithmic systems is their ability to manage risk. Can algorithms make mistakes? Absolutely, and they do at times. However, in terms of understanding as much about the “knowable” risks involved with an investment, algorithms have shown to be incredibly effective if properly engineered. Humans are notorious for having emotional biases and breaking systematic rules. Algorithms don’t have personal issues or take sick days, they are the definition of consistency.

Decoding Wall St

Source: Decoding Wall St — Infographic / QuantConnect

Depending on the firm and their objectives, they’ll employ algorithms to either perform part or all of their portfolio management tasks. It is not uncommon for a firm to use algorithms to discover favorable investment opportunities and then have the investment manager make the final decision. Other firms take a “black box” approach and allow the algorithms to make each and every decision start to finish. Most of these “black box” firms have algorithmic models that are so proficient it is actually safest to leave all tasks to the algorithms rather than use human discretion.

What happens if things go awry?

Decoding Wall St

Source: Decoding Wall St — Infographic / QuantConnect

There are a few instances in the last 15 years that algorithms have been blamed for sudden, unexpected crashes in the market. Like most systems, especially those of human-design, algorithms are not perfect. According to Bloomberg data, in 2014 the daily value of transactions on the US stock market averages $279 billion. With billions of shares being traded daily, it is surprising catastrophic crashes don’t happen more often. Market regulators and exchanges put in place rules and circuit-breakers to help mitigate the risk of future large-scale flash crashes. A group by the name of Nanex monitors streaming market data on a nanosecond (1 billionth of a second) timescale and is able to see when such events take place. According to research, there are numerous mini flash crashes every day in the market. Is this surprising? Yes. However, it is nothing to be alarmed by. Systems of this complexity are expected to experience small hiccups. Just as the human body is constantly dealing with infections and cancerous cells, so do the risk management systems deal with errors in the market. Occasionally they fail and you get events like those shown above. The key to further mitigating these risks is by gaining information from these small incidents in order to prevent future large scale incidents.

In the right hands, algorithms are powerful and effective.

So why do most financial advisors not convert to algorithmic investing?

There are multiple answers to this question, but the simple reason for most: they can’t. Remember back to the beginning of this article the job description for a position as a Quantitative Analyst? The background necessary to design such systems is drastically different than that of most financial advisors who often have degrees in business or finance. While a PhD or Masters is not essential to be successful in modeling algorithms, a solid understanding of math and computer programing is. Most financial advisors simply do not have this sort of background and will not be going back to school to attain it. What they’ve been doing for years "works" for them, so why change it? If asked, the majority of advisors will say they simply do not understand the models. A minority of them may say they understand them, or even that they use some of the models, but in reality their scope of understand is often quite limited. Additionally, most large investment firms and banks have policies against their advisors using algorithmic models to invest, and for good reason. As with any strategy, there are certain risks, and with such models these risks can be complex and difficult to understand. In the hands of someone inexperienced, the investment outcomes can be anything but favorable. The writing is on the walls, however, and in the coming years many of the investment practices advisors have relied on for decades to manage their clients’ portfolios will become obsolete. In a world dominated by technology and rapid advancement, many advisors will be forced to adapt or they too will be replaced.

But human analysts do a good job, why replace them with algorithms?

While most analysts are good at what they do, they too can be plagued with biases or corporate agendas. The infographic below helps provide a peek into what make Wall Street analysts tick. Some of these numbers may come to you as a surprise. This information shows it is not uncommon for advisors to rely on misinformation when it comes to their investment decisions if their research is partially or fully based on the opinions of analysts. Whether it is an archaic investment strategy or the opinion of a specific analyst, there may be hidden issues and risks being passed on to clients. Just 13% of analysts said retail clients are their most important clients. Not good for the small guy!

“Financial analysts often walk a ‘very fine line’ in balancing their responsibilities to serve their clients, maintain relationships with corporate managers, and follow the law." - McCombs Accounting Professor Michael Clement


Data by Michael Clement

Infographic by Chris Philpot

Want a deeper look?

For those interested in knowing more about financial algorithms, below is some information on the main types use in the markets today.

Decoding Wall St

Source: Decoding Wall St — Infographic / QuantConnect

The deepest look into the market

Curious what high frequency algorithmic trading looks like in the market?

Below is animation from a few seconds on November 7 2012 compiled by Nanex from data they collect on every transaction in the market.

HFT trades

Source: Radiolab

As you can see, when it comes to visualizing what actually goes on inside the market the true complexity of the system is obvious. Thousands of trade orders take place in the amount of time it takes to blink your eye (literally!). The financial markets have come a long way in the past decade and it is time investors begin to understand the implications of these new advancements.

Is your investment approach keeping up?

Just as you want your doctor to remain informed on advances in medical technology when addressing your health, you should also want your advisor to keep up with advances in the financial world. In this industry it is important to seek continued education and access to new technology in order to make more informed investment decisions. Whether it is your retirement account or brokerage account, small improvements now can equal large financial benefits when compounded over multiple decades.

Additional sources:

Decoding Wall St QuantConnect

Go Top
comments powered by Disqus