Welcome to the sixth issue of the Infinitary Connections, the official Substack of the Infinitary Fund. This is a weekly newsletter where Nic and I will be sharing some of our insights across the world of finance, economics, technology, and more. As partners in the Infinitary Fund, we have a unique perspective on the world, the markets, and the shape of things to come.
This week we discuss quant funds. If you aren’t subscribed, you can subscribe for free below.
And make sure to like this post so we know if it resonated.
Let’s dive in.
BIG NEWS
Before we start, we are pleased to announce that Mike will be speaking at the prestigious QuantMinds International Conference in London this November. This is the world’s leading quant finance event and Mike will be sharing the stages with some of the top minds, asset managers, and academics in the world from companies such as BlackRock, JP Morgan, BNP Paribas, NatWest, Amundi, and Morgan Stanley.
QUANT 101
A quantitative fund, also known as a quant fund, is an investment fund whose securities are chosen based on numerical data compiled through quantitative analysis. The strategy relies on mathematical models and algorithms to make investment decisions. These models and techniques are used to identify profitable trading opportunities and manage risk These funds are considered non-traditional and passive.
As quant funds rely on algorithmic or systematically programmed investment strategies, they don’t require actual fund managers. And by fund managers, we mean human beings with experiences, judgment, or opinions. They are no longer part of the investment process and many believe that this removes the possibility of risks and losses associated with human error.
DATA IS KEY
The foundation of every major quant fund stems from the available data. Large swathes of financial data is collected from a multitude of sources. This includes the latest economic indicators, company financials, social media activity, news sentiment and company financials. This data is processed in order to remove outliers, errors, and inconsistencies.
Data allows for the actual programming of the fund. Quant funds rely on thousands of trading signals that they rely on to make decisions such as real-time company news to trending global asset values. The data is also used for backtesting. Historical data is used to simulate the performance of the fund’s strategy over time. This process helps in understanding how the strategy would have performed in the past and reveals potential weaknesses or limitations.
Data is also used for building a sophisticated model based on quality, value, financial strength, and momentum via a proprietary algorithm (we have our unique algorithm that Nic developed). Speaking of Nic, quant funds need highly intelligent and diligent analysts…or as some call them, quant jocks.
QUANT JOCKS
Quantitative analysts, often called quants or quant jocks, are crucial to a quant fund. They play a major role in the model development. These analysts are highly skilled in mathematics, statistics, and computer programming. They design and create sophisticated mathematical models to analyze historical data and identify patterns that can be used to predict future market movements. Nic, for example, has a degree in mathematics and also worked as an adjunct instructor of finite mathematics. His mathematical prowess gives our fund its competitive advantage.
The most famous quant jock is James Harris Simons, a mathematician and the founder of Renaissance Technologies. He has been called the greatest investor on Wall Street and the most successful hedge fund manager of all time.
RISKS AND BLACK SWAN EVENTS
Quant funds are not without risk. When they fail, they fail significantly. This is often due to the simple fact that quant funds in general are based on historical events and the past doesn't always repeat itself in the future.
Sophisticated quant managers will be constantly adding new aspects to their models in order to predict future events, it is simply impossible to predict the future every time. Circumstances are constantly shifting and volatility always plays a role. Quant funds can become overwhelmed when the economy and markets are experiencing greater than average volatility. Additionally, these funds can struggle to account for unprecedented events, often referred to as "black swan" events.
A black what?
A black swan is defined as "an unpredictable event that is beyond what is normally expected of a situation and has potentially severe consequences. Black swan events are characterized by their extreme rarity, severe impact, and the widespread insistence they were obvious in hindsight."
Examples of black swan events include the September 11th attacks, the 2008 financial crisis, and the Covid 19 pandemic. Such events can result in extreme market movements that deviate from the models' predictions.
Quant funds can also pose a risk when they are marketed as bear-proof or are based on short strategies, which is highly risky. Predicting downturns using derivatives and combining leverage can be dangerous. One wrong decision can implode the entire fund. Due to this, quant funds must continuously monitor their positions and the performance of their strategies in real-time. If the market dynamics change or the strategies show signs of underperformance, adjustments and refinements need to be made promptly.
OUR FUND STRATEGY
Our approach to investing is 99.99% a numbers-in, numbers-out process. This means that, for the most part, every decision with an investor's money is based on an algorithm, and this algorithm does not change over time, except at a future date, at which time we will make an announcement beforehand.
First, investor money enters our bank account and we transfer it to our Vanguard account. From the moment we accept the money, we know that we will be splitting up the accepted money equally between five companies in the current S&P 500 list. Thus, the starting number of stocks and allocation amounts are both predetermined quantitatively (in order to reduce the uncertainty of an expected result from our system).
Although our implementation is simple in the sense that we buy-and-hold an equal allocation of five stocks, our analysis process is very extensive. Our algorithm relies just on the price data of stocks. Price data is very challenging with which to make money because it is so freely available and one is competing against the entire world, which may include people with better resources and opportunities. But we have noticed (even before the AI boom of 2023) that too much of competitive machine-learning involves an overcrowding of methods. What we have produced is unique in the way we model our analytical process. This gives us a way to make returns which are different from the crowd.
Our algorithm looks for small disturbances in the stock prices of large companies, which are likely to turn into something larger and different later in time. In other words, we are predicting the next hurricane season in Florida based on a butterfly's wings in Tokyo. Why do we do it this way(?), because the BIG returns are in knowing not the momentum of the past, but the axiomatic makeup of the future market regime. We don't ride the current waves, we aim for the next big wave that nobody can see yet. This method is however a potential slippery slope to disaster if one is not armed with the proper mathematics.
Our algorithm is based on a model of how structural novelty happens and takes inspiration from complex adaptive systems. Most of the calculations that are performed by the algorithm are to prevent overfitting (overfitting means finding a solution that is too-good-to-be-true by learning too much about the past instead of the future).We order our stocks according to a metric from which we build a “Bridge” by selecting the best five stocks.At this point, we perform the 0.01% part (only approximately) which is non-algorithmic. We read the fundamentals and company news of the stocks that we want to buy. If anything seems suspicious to us (in terms of new spinoff companies or near-bankruptcies for instance) then we remove it from our top five stocks.
Then we divide the money equally into five parts and buy the five stocks, and then hold them for one year. At the one year mark, we re-perform the algorithm and update the money using a new set of five stocks (which at times may include stocks from the previous year(s)).We believe in sticking to our guns more than relying on agility. The stock market does not work linearly so we don’t expect it to do so. We give our stocks the time they need to grow. Market downturns are events for which we have extensively back-tested our system. Our system was designed to recover extremely quickly after market downturns, and we never use leverage in our portfolio.
Our research process to improve ourselves involves making our algorithm less susceptible to downturns and more susceptible to upsurges, but it is ALL numeric in nature and implemented into our algorithm (...we don’t shoot from the hip).
Thus, our method is to use a mathematical function on behalf of our partners which creates extremely competitive returns by using a computational definition of what constitutes a high-growth stock. So despite the fact that we are interested in qualities, we are relying on numeric quantities to make our decisions. For this reason, we call ourselves a mathematical investing aggressive growth fund. We are researching the most advanced math in the world in order to use the knowledge that results to compute the best investments for our partners, while also catalyzing an industry of pure mathematics. So for us, we are a quantitative fund because we are 99.99% relying on an emotionless mathematical function of the stock prices for the S&P 500.
If you are interested in learning more, please reach out. Don’t forget to share this post as it truly helps us.
We will see you next time.
Mike and Nic.