What's wrong with Quant Investing? Let's try to challenge it

Every Finance or Stock Market Guy with little understanding on Machine Learning and Artificial Intelligence would start dreaming about creating a algorithm to predict the stock market. Of course, mine also started the same way. Let me first explain the basics around the quants to better understand the drawbacks and how I try to challenge the very drawbacks of the quants.

What is  Quant?

I mean Quantitative Investing. So, What is Quantitative Investment Strategy?

In this strategy the model of valuation would be created by utilizing the skills of the Finance, Stats and Coding to evaluate a security (shares) and predict the pricing of the same.

Investopedia says "Quantitative investment strategies have evolved into complex tools with the advent of modern computers but the strategies' roots go back over 80 years. They are typically run by highly educated teams and use proprietary models to increase their ability to beat the market. "

Why Quants?

A Person need to stick on to a fewer stocks where a algorithmic model would be able to analyses all stocks at a lesser cost and lesser time. But creating one such model would be highly costly unless willing to learn all three 😂 and DIY.

What's Impacting the Quants?

These Quants are majorly developed by large institutional investors. When large institutional investors use these quants and invest, market correction is very immediate. Economic Factors are completely ignored by the quants at a major stake.

Now, Let's see how the economic factors were being ignored

The economic projections are expected to be part of the model but all the time it's just the historical data that is getting into the model. The real time political factors, climate change and mostly the psychology factors are often ignored by the model

Let's get into the reason Why?

The difference in understanding the non - economic factors and the volatility in those factors is the major reason, on the failure of the quant model. Back in the end of 20th Century when the quant models are initially started up, the ability of the technologies hasn't progressed as it is today. The constant predictions would require big data.

What's this Big Data?

Huge Data, both structured and unstructured. Everywhere the definition is ambiguous, I will take an example to explain a Big Data Scenario. Databases where it includes "Weather Data, Local Economy Cash Flows and along with Accounting books"

What might go wrong in testing?

When model is analyzing only historical market data, that it is quite obvious that tomorrowcan't be as tomorrow. So, there is a higher chance for the model notwithstanding the market accuracy test at a desired rate. But it is also accurate in a scenario based tests performed on the model. A well documented testing of the model would certainly prove it's worth.

  

My Interaction

On my interaction with the founders who are aspiring to create a quant model, makes me feel the deep tech growth happening in India. Even though the predictability is debatable, but there is always a chance for the finer results. At the current level of, the ability of accurate prediction is higher at the short term level and for the long term human intervention is a must have.

What's the present scenario?

A Quant model assisted by Human is the current possibility. Of course, it won't be able to reduce the costs that are paid to analysts but the faster analysis would help the investors to profit the ever changing valuations of the market.