Of recent interest in financial markets is the desire to analyze the implications ofscores vs. factors and their implications on financial markets - in particular their comparativeusefulness for investing. By factor, we mean directly observable numeric quantities such asunemployment rates, yield rates, GDP, etc. By scores, we mean the output of a model on a nonnumeric quantity, such as an ESG score utilizing twitter text data. The idea behind this project is toformalize a comparative mathematical theory between these two entities.
A wise institutional investor should diversify his investment into severaldifferent asset classes, but what asset classes he should consider in the first place is a non-trivialdecision. We therefore have proposed several different metrics for understanding the interdependencies between assets so that the question of if a collection of assets forms an asset classbecomes a rigorous clustering problem. We’d also be principally interested in if cryptocurrencieswould satisfy such a definition.