2 min read

How does an investment thesis enable a new approach for custom indexing?

How does an investment thesis enable a new approach for custom indexing?
Photo by DeepMind on Unsplash

Noonum provides a solution useful for custom indexing with a new way of digitizing an investment strategy.  While this blog will be mostly about technology, first let us understand the problem that technology is trying to solve.  So, what is the problem with custom indexing and how will an AI-based approach address it?

The investment world primarily uses the approach of ‘factor’ investing for custom indexing.  Factor investing is simply a way of saying that an investment strategy can be optimized for performance by taking into account various properties of a security and using appropriate ‘mixes’ of securities based on those properties.  For example, ‘growth’ or ‘value’ strategies look at the expected revenue and earnings growth of available securities. Momentum strategies increase exposure to securities that have recently performed well, while a myriad of other factors can be incorporated, from market cap, to geography, to liquidity. Once a portfolio is built, an investment manager rebalances the portfolio periodically, changing security weights and holdings to reflect the strategy’s desired factor exposure.

Technology for factor investing includes creating and using the categories or properties of securities, like price momentum or growth classification, as well as the ‘weighting’ or value that describes the factor’s strength. The methodology underlying these factors is generally consistent across the investing community. While the precise criteria and thresholds might differ from one to the next, an asset manager’s classification of a security as defensive or momentum-based is generally consistent.

But traditional factor investing is a saturated and highly competitive space within the asset management world. And there's a huge gap in the market for innovative products that combine an asset manager’s portfolio expertise with the ability to incorporate new factors into the custom index process. Natural language-derived thematic investing has unprecedented potential to offer a new way to build investment solutions at scale, with proven performance and defensible strategies.

Noonum solves this problem by providing a proprietary technological approach to custom indexing that allows flexibility to meet different investor needs and generates a superior ability to understand securities. The starting point for any new investment product is the portfolio manager’s thesis – what factor or aspect of the market will the product track? We call this strategy building and help the user understand and evaluate all of the components to build their thesis.

Let’s look explore this through a sample thesis – Artificial Intelligence.

There are many ways to express Artificial Intelligence. Noonum enables a user to find and evaluate each of the different inputs to the thesis: data centers providing the computing power, developers of machine learning and neural network tools, end consumers who benefit from the technology, and robotics and shipping logistics.  A user needs to understand how each of these themes act as inputs into the overall strategy, how each theme has performed and evolved, and what securities provide exposure to each theme.

The rigor and methodology behind the strategy inputs and the ability to quickly test and analyze the index has been absent from thematic portfolio construction. Through customized ML models and proprietary data processing capabilities, Noonum’s core IP is in how we map and translate the investment thesis into exposure scores that represent how material the thesis is to a given security. These scores can then be used to weight the individual securities within a custom index. With two awarded patents, we believe our unique approach will power the next generation of custom index investing. Noonum brings the potential of AI technology to investing.