Predictions for the future can seem fantastical – that is, until they come true. Sure, we may not be flying around in hover cars just yet, but wrist phones, electric vehicles and wireless headphones all seemed like impossibilities at one time too. All that is to say, technology can advance pretty quickly. Forward thinking asset managers are using these rapid advances in technology to make their lives easier by implementing artificial intelligence techniques within their portfolios. Here, we’ll walk through some of the ways investment managers are succeeding with AI and machine learning using our platform – Boosted Insights – and some quick wins that other institutional investors can action with AI and ML.
How can AI and machine learning help me generate ideas in my portfolio?
Think of Boosted Insights’ AI and machine learning as a hyper-advanced data sorter – it takes complex inputs like stock price data, alternative data, macroeconomic data, apply the variables an investment manager thinks are important in stock picking, learn from it, and generate a predictive and constantly refreshed, ranked list of stocks within the manager’s universe.
From that list, the manager can see which stocks the machine “likes” and “dislikes” – those that move up or down significantly in the dynamically ranked list. Those serve as flags for the asset manager to look more deeply into an individual stock. We highlight these flags on our portfolio one-pagers, which you can read more about here.
To work at its fullest potential, our AI needs an institutional investor’s expertise, both for creating the initial model to flag opportunities, and to look closer into those potential opportunities. Machine learning is good at highlighting anomalies or possible opportunities thanks to its massive processing and learning power. Where the manager’s financial experience comes in is deciding which of those flags are insights to put to action within their equity portfolios. We like to think that the relationship of AI plus human insight is greater than the sum of its parts.
How can artificial intelligence and machine learning help me lower risk?
AI helps lower risk for institutional investors by isolating potential risks within their equity portfolios. The biggest benefit of machine learning is its data-driven ability to learn and adapt. As it applies to equity capital markets, good AI can help investors identify sectors or individual stocks that are susceptible to risk. Within Boosted Insights, we have a number of ways to help protect asset managers from risk, like Semantic Risk Factors, Similar Stocks and Factor Constraints – we’ll walk through those options below.
Can I use NLP to improve my portfolio?
Semantic risk factors use natural language processing to isolate industries that move in opposite correlation with one another – stocks in risk group A are expected to move converse to stocks in risk group B. This can help give managers early signals for industries to watch out for. Read more about semantic risk factors here.
How can I use AI to create intelligent hedges?
Similar stocks uses AI to help managers build smart hedge baskets. Like when it is surfacing portfolio opportunities, the super power of AI and machine learning is its ability to learn from vast amounts of big data. Similar stocks combines stock price data with machine learned data to create correlations and similarity scores for individual stocks or baskets of stocks. This allows investment managers to curate smart and adaptive hedge baskets within their portfolios. Read more about AI-powered hedge baskets here.
Can I use AI to improve factor timing in my portfolio?
Another way portfolio managers can limit risk within their book is by implementing factor constraints. We have found evidence that using machine learning factor constraints to limit portfolio exposure has lessened sharp drawdowns. With these factor constraints, an asset manager is able to limit their exposure to severe risk by keeping allocations to individual risk factors restricted. Our ML factor constraints have proven to be especially helpful during regime shifts, like when sentiment shifts from momentum to fundamental stocks, or when meme type stocks (like GME) can blow up portfolios. Here’s a quote from our post on avoiding short blow-ups:
“In fact, taking the average performance of the Machine 8 bucket from January 11th to January 27th is 59.89%, even removing GME and AMC from the list the average performance is still 11.40%. The key thing is that our machine learning was able to identify this basket, highlighting it as the 8th largest risk in the market (at the time). It didn’t know that this was going to happen – but using it in portfolio construction would’ve allowed it to help contain the damage.”
How quickly can I get started with AI?
One thing we hear over and over from our asset manager clients (not to mention employees from the institutional equity space) is that investment managers are time constrained and don’t have capacity to institute entirely new procedures and processes. Quick AI wins that lead to long term change are instrumental to implementing and adopting artificial intelligence for these folks.
As these users can be extremely time constrained, the idea of building an AI prediction system themselves can be too time consuming and too resource intensive. They know they need to adopt AI (machine learning is altering every industry – the most data-driven asset managers will see the most success going forward), but don’t have time to build it themselves. For these teams, an over-the-counter suite of machine learning tools is the best way to see early success.
Boosted Insights is designed specifically for capital markets prediction. Our team members consist of ex portfolio managers, equity salespeople, and machine learning and data science experts. Our customer success teams have helped investment managers score quick wins with AI within as little as a month (we will note that our clients that have seen the most success on Boosted Insights realize that the best results come from a collaborative process between our CS team, and their investment managers in achieving their investment mandates/goals). Fundamental and quantitative users have seen the value in Boosted Insights, both as a big data-driven idea generation platform and as an advanced lab to test data and signals (and one that provides uncorrelated signal).
This is only a brief description of some of the ways forward thinking asset managers are harnessing the power of AI. There are any number of use cases institutional investors can take advantage of with Boosted Insights, including idea generation and dynamic stock screening, smart hedge baskets, portfolio optimization, evaluating alternative data and signals, and factor analysis. If you’re curious how to get started with AI and machine learning within your organization, please reach out to us here.