We have launched a new portfolio – the Global Multi-Strategy – which has been built using a cutting-edge quantitative approach that we believe can navigate the uncertain years ahead. The portfolio’s primary aim is to generate returns across market regimes by combining multiple sub-strategies across styles and asset classes.
The Global Multi-Strategy portfolio has been designed in partnership with Second Foundation Partners, the publisher of Epsilon Theory and manager of a global macro hedge fund that utilizes their pioneering research. This research focuses on generating market-based returns by making trading decisions using natural language processing (NLP) analysis of financial news.
Together, we have built a sophisticated quantitative strategy, implemented via liquid ETFs, that Save’s Market Savings customers can choose for their next FDIC-insured deposit.
Approach
Whether markets go up, down, or sideways, sources of returns exist. Extracting those returns does not require you to personally have the right opinion about markets themselves; instead, what arguably matters most in the short to medium term is figuring out what everyone else thinks and what they, en masse, may do next.
As the poker saying goes, “you don’t just play the cards, you play the players”. The fundamentals of a company or sector (the cards) are important, but even more important is figuring out how other market participants (the players) will react to news and rumors about those fundamentals.
For example, you might think (at the time of writing in December 2022) that stocks are still expensive and “should” fall further – and you could be right sometime in 2023 – but if the market decides otherwise in January (responding positively to, say, news about China re-opening or US jobs), then your investment portfolio may suffer.
This is the thinking that forms the basis of our new Global Multi-Strategy portfolio, designed in partnership with Second Foundation Partners and benefiting from a decade of quantitative research and development into how the themes and patterns found in financial media – called “narratives” – interact with financial markets and their participants.
Portfolio Composition
The portfolio contains 6 sub-strategies:
- Equity Beta
- Fixed Income
- Currencies
- Commodity Beta
- Commodity Relative Value*
- U.S. Equity Sectors Relative Value*
*where “relative value” refers to a long-short approach that takes long positions in some preferred assets, versus short positions in some less favored assets, seeking to gain from the relative outperformance of the preferred assets.
Each sub-strategy utilizes a different set of trading signals, with the aim that by combining the different sub-strategies, the portfolio is potentially able to generate returns across all market environments.
Signal Calculation
Using both Big Data (millions of media articles and transcripts) and Big Compute (trillions of unstructured data operations), the strategy first analyzes the language across everything published in the English language about companies and markets:
Narrative maps, like the example below, are then produced and their evolution is analyzed. Each node on the map is an article, and the physical distance between any two articles indicates how linguistically similar or different those articles are.
A numerically large cluster of closely spaced nodes suggests a lot of focus on a given topic (say, inflation), and that those articles are largely in agreement in their assessment of that topic (let’s say, “inflation is high”). This also allows us to assess the stage in the life-cycle of the inflation theme, as well as its relative importance compared to other narratives, for example – is inflation currently emerging as an important topic grabbing a lot of new attention, or is its significance waning, giving way to ‘recession fears’ as the predominant topic?
In other words, the mathematical operations search for specific narratives that spark specific investor behaviors in response. These behavioral “triggers” are systematically evaluated daily and drive dozens of individual buy-and-sell signals across dozens of global assets. Once triggered, signals tend to have a persistent lifespan of the order of weeks.
These buy-and-sell signals are systematically rolled up into the 6 sub-strategies of the Global Multi-Strategy portfolio.
Current Signals (Dec 2022)
At the time of writing, the analysis suggests that: financial media believes inflation is declining, the outlook for interest rates is shifting toward easing (reducing), and a recession could be approaching; additionally, the analysis shows there has been significant recent bearish thinking, and that common knowledge of the bullish case for energy (relative to other sectors) has reached saturation, among other things.
Final Comment – Play the players
By expeditiously and accurately capturing the narrative states relevant to market participants (the players), and skillfully determining what those players might subsequently be predisposed to do next, a narrative-based approach can arguably trigger trading decisions faster, and with a greater success rate, than other traditional investment approaches that only consider fundamentals (the cards).
Subsequently, we believe the Global Multi-Strategy portfolio can deliver the return profile that our customers are looking for.
