Artificial Intelligence and Unstructured Data are Fundamentally Changing the Way We Will Pick Stocks

How AI is Transforming Investment Research

• Leaps in computer processing abilities, allowing the cost of quality processing tech to decrease.
• Increase in the availability of accessible data, which AI can use as needed.
• With cloud capabilities, the cost of storing decreased
• AI for investment guidance and data analysis is now cost-effective enough to become a business model
• The exponential growth of unstructured data

Context/ Background

AI can provide up-to-date information on product movement throughout the supply chain, analysis of weather forecasts, and outline search engine topics,

How did we get there?

• Research studies have proven the validity of the use of Neural network to predict stock performance
• Natural Language Processing is now used for the analysis of earning calls, companies like Aiera,
• An S&P paper suggests that the aggregate sentiment of analysts historically enhanced the predictability of the 3-month FY1 EPS analyst revision signal — S&P
• The same paper infers that Firms That Referenced the Most Positive Descriptors around Their Financials Outperformed Historically. Firms whose executives most frequently articulated references to growth- and expansion-related descriptors around I) revenue II) earnings or III) profitability topics outperformed their counterparts by 9.16%, 8.60% and 6.76% per year, respectively (Table 1). See S&P report — Applying NLP Using Domain Knowledge to Capture Alpha from Transcripts

Previously noted

• Artificial intelligence goes further by enabling the system to adapt based on the information it receives. At Man, engineers set parameters: exposure caps, asset class, volatility, trading costs, etc. Compliance and risk management rules are ingrained into the system’s DNA, preventing it from going rogue or breaking the law as a fast track to profit. — Bloomberg
• Man Group Plc is expanding the use of algorithms that evaluate risk, pricing and timing in financial markets and learn from their mistakes — Bloomberg
• There are now a number of players in the space

Be smart

• Companies like Boosted.ai have developed solutions like Boosted Insights that continuously learn from your inputs to predictively rank stocks. It discovers and explains which combination of features is important in different periods
• Sentifi machines will go through millions of data pieces to tell you which companies are being discussed the most and the topics around which discussions are centered.
• Clarity AI uses algorithms to provide forward-looking analysis of impactacross more than 30,000 companies who have issued equity or debt, spanning almost 200 countries.
• The AI-Powered ETF (AI EQ) invests primarily in US exchange-listed equity securities based on the results of a quantitative model which identifies approximately 30 to 70 companies with the greatest potential over the next 12 months for appreciation and their corresponding weights.
• EQBot, the company behind the AIEQ ETF is an interesting case:

And they use the following data sources

• Sigmoidal designed and developed an AI-based tool to track and analyze clients’ sentiment towards companies from their portfolio, working in real-time. It also automatically generates high-level reports, including only the crucial information.
• According to Institutional Investor, AI-led hedge funds produced cumulative returns of 34 percent in the three years through May, a report Tuesday from consulting and research firm Cerulli shows.

Why it matters

• Companies like Kavout, enhanced by AI and machine learning technologies, delivers risk-adjusted alpha signals as a data feed and portfolio design to equity buy-side firms.
• Companies like Trade Ideas make artificial intelligence (AI) accessible to professional traders

Driving the news

• Israeli fintech company TipRanks received a $77 million investment, led by Prytek and More Investment House. Using Natural Language Processing it tracks and measures the performance of over 7,000 professional analysts and enables investors to see their track record and evaluate their advice.
• Carlyle invested in a French AI startup after a year as a client, in Sesamm, a developer of artificial intelligence-powered analytical software used by investment managers and banks to crunch through large data sets and analyze reams of text to identify investment opportunities.

Yes but…

• According to the WSJ, AI-based investing strategies have struggled to live up to some of the more inflated expectations for their performance.
• PwC warns of the lack of transparency as algorithms learn and become more complex
• Most of the machine-learning algorithms used in finance are supervised, meaning that the model learns to recognize patterns by analyzing historical examples. With the pandemic-led global lockdown being a new and unforeseeable event, it is extremely difficult for the models to adapt to the various scenarios dynamically, explains Cerulli.

By the numbers

• The global AI in Fintech market was estimated at USD 6.67 billion in 2019 and is expected to reach USD 22.6 billion by 2025.
• An analysis by Cerulli Associates of the assets under management (AUM) and net new flows of Europe-domiciled AI-led funds from 2013 to April this year shows strong AUM growth from 2016 to 2019. The cumulative return of AI-led hedge funds was almost three times higher than that of the overall hedge fund universe during this period: 33.9% compared to 12.1%. Despite this, AI-led hedge funds’ net new flows fell slightly last year, before dropping sharply between January and April.

What people say

• These results don’t mean that AI is worthless, Prof. Avramov is quick to add. “It’s just that its potential has yet to be proven,” he says. “AI definitely has promise, perhaps not just as much promise as some have made it out to appear.” WSJ
• “There has long been suspicion of the ability of AI to react to unexpected events, such as the coronavirus pandemic, but there is now a sense that the technology has advanced to the point where it is better able to adapt to unforeseen scenarios via the ever-growing amount of market data available,” says Justina Deveikyte, associate director, European institutional research at Cerulli.
• Renowned investor Paul Tudor Jones once remarked, “No human is better than a machine, but no machine is better than a human with a machine.”
• PanAgora’s director of equity investments Mike Chen told Cerulli that constructing machine-learning factors is a “balancing act” as fund managers don’t want them to react to quickly to “noise in the market information” — or so slowly that they miss a trend, according to the report.

What I think

• Investment research is the perfect use for AI capabilities
• Equity research deals with a lot of data
• Investment Managers have to have a data analytics strategy involving AI this is now a must-have, not a nice to have anymore
• Researchers are building predicting models like this one, that seems extremely promising: The TEI@I Methodology Framework for Crude Oil Price Forecasting

Go deeper

What to listen

Leverage Artificial Intelligence in Your Investment Strategy with David Aferiat

What to read

• The Most Powerful Artificial Intelligence Knows Nothing About Investing. That’s Perfectly Okay. Indeed, that’s the point — Institutional Investor
  (Must Read)
• Read Artificial in Asset Management by the CFA Institute

This post is part of Convergences by Melvine. A series exploring how software is changing every corner of human activities. Melvine Manchau