Welcome to the Third Wave of Investing — Machine Intelligence
The history of investing has often been a story of incremental improvements. Things are about to get more interesting quickly.
First Wave: Fundamental Research
The principles of fundamental stock analysis established by the likes of Benjamin Graham and Warren Buffett have defined investing for nearly a century. Although the tools have become more sophisticated, the idea is the same: by studying a company’s financial statements and other industry and economic metrics, investors can form a deep understanding of the future return prospects of a company to construct prudent portfolios.
Second Wave: Quantitative, Rules-based Investing
Advancements in mathematical modeling allowed investors to systematically analyze a growing deluge of data to access factor premiums and other market anomalies through quantitative methods. These solutions are scalable, efficient and most importantly, not subject to the vagaries of human emotion. But they fall short in capturing the full qualitative aspects of a company and can’t absorb new information without being reprogramed, making them best suited for capturing one or a few well-known market premiums that often go in and out of favor.
Third Wave: Human Plus Machine
What if you could combine the depth and nuance of fundamental research, add the massive data processing capabilities of traditional quant strategies, and turbocharge it by unshackling computers from preset rules, giving them the power to learn and make deductions faster and more efficiently than the human brain? We call this machine intelligence (MI)—applying machine learning to stock selection, enhanced by human insight and traditional quantitative methods.
Maximize alpha generation by blending human intuition with quantitative insight and machine intelligence.
Machine Intelligence in Practice
An AI-driven investing platform needs a few things to work well:
- A large set of clean data: The Voya Machine Intelligence (VMI) team has spent thousands of hours scrubbing and standardizing a broad span of data (similar to how fundamental analysts look at company data) in order to determine materiality and to curate factors into features.
- The rules of the game: The VMI team has trained 26 virtual analysts and 45 virtual traders on the rules of the investing game to uncover persistent patterns in company fundamental data that drive outperformance. But we don’t tell them how to learn—we arm them with the right tools and allow them to learn on their own as markets evolve.
- Human partners for risk management and trading: Humans review buy and sell decisions and execute all trades to ensure the system is abiding its constraints and to account for pre-market announcements.
This approach brings humans and machines together in a symbiotic relationship, driven by a goal to create portfolios devoid of destructive notions of fear and greed. In combining the forces of humans and machines, we seek to create portfolios free of the destructive impulses of fear and greed. Moreover, this approach is highly adaptable, making it well-suited for crafting custom portfolios that target specific objectives and preferences, such as sustainable investing mandates.
“As institutions consider opportunities to invest IN AI, the next step may be to invest WITH AI.” - Gareth Shepherd, PhD, CFA, Co-Head Voya Machine Intelligence (VMI), Portfolio Manager