New England Nautical Geography? A Quantitative Analysis of Hedge Fund Names

Late last year, two academics published an article associating hedge fund capital flows with name ‘gravitas’. They defined gravitas as “a combination of words from geopolitics and economics… or suggesting power”. Their research validated an old joke in the hedge fund industry that all funds name themselves after trees, classical figures, or geology.

SumZero quantitatively explored the validity of this across the 5,000+ hedge funds listed on our platform. To do so, our data team implemented natural language processing techniques to cluster trends in naming convention. In addition to some very prominent topical themes, there are numerous granular sub-categories which can be explored in an interactive map (e.g. "New England locations", "nautical", “Hindi Words”) below.

Link to SumZero Word Chart

Analysis / Methodology

We found that the most common root words used in a hedge fund name (after stripping common suffixes like capital, capital management, asset management, partners, capital partners, etc.) were the following:

'value', 'street', 'point', 'river', 'hill', 'rock', 'view', 'pacific',
'bridge', 'blue', 'wealth', 'new', 'stone',
'wood', 'park', 'global', 'first', 'silver', 'tree',
'peak', 'lane', 'creek', 'star', 'alpha', 'oak', 'american', 'green',
‘asia', 'north', 'west', 'bay'

The average length of the suffix-less root is 1.6 words. The most typical hedge fund name ends up being something like ‘Value Street Capital’.

As to category, we found six large groupings:

1. Names/Places
2. Adjectives/Abstract Terms
3. Nature
4. Business
5. Science/Tech
6. Mythology/History

A caveat to this analysis - we did not include words that aren’t in the English dictionary, which excludes from our analysis Uncommon People/Place Names (Balyasny, Coatue, Leucadia), Acronyms (AQR, SAC, PDT), and Latin Sounding (but not Latin) words (Visum, Diversis, Infinitas).

Vectorized Root Names

Next we took the root words and "vectorized" them using google's word2vec system. "Vectorizing" words is a way to represent words as points in space where similar words are nearer to each other. For instance using the root fund name "fidelity", we find the most similar terms are:

faithfulness
premarital chastity
compatibility
intelligibility
lossless compression
intimacy
functional equivalence
monogamy

This gives us a way to map words by meaning, though as you can see in the visualization, the meanings are sometimes unexpected. E.g. most similar word to "Marble Arch" is "fluted column", but the most similar word to "Soros" is "Glenn Beck". Putting our list of hedge fund root names into this "vectorized" format we generated a two dimensional map of the world of hedge fund names grouped into a few major categories as shown in the image below. There are many many surprising subcategories, and sub-sub categories such as "Alcoholic Beverages", "Formal Meetings", and "Scientific Instruments".