Cooking Up the Ingredients of an Investment Recipe

A little-noticed 2014 working paper by Boris Gnedenko and Igor Yelnik carried the intriguing title “Hypercube in the Kitchen: Reading a Menu of Active Investment Strategies.”  It proposed a method of “skill-based classification . . . to assist investors in better understanding a menu of available investment strategies as well as to help asset managers to position themselves on that menu.”

The “hypercube” in the title referenced the combination of five dimensions the authors used to characterize an asset manager’s investment approach — the “kitchen” signified the manager’s concerted effort to cook up a particular recipe of attributes that delivers tasty results for its clients.

The classification problem

A challenge with any attempt to classify a manager (or for a manager to classify and market itself) is that a firm can “implement dozens of investment processes simultaneously,” so the recipe is hard to explain or defend — and hard to decipher.  There’s more than a few ingredients.

No classification scheme is complete or final.  Boundaries move over time and new categories are created, posing a challenge for all involved, since where the lines are drawn affects the behavior of managers and allocators alike.  The dividing lines migrate, be they for asset classes or industry sectors or hedge fund strategy types or anything else.

Skills

The authors sought out a way of classifying managers that could augment existing methods of slotting them.  They proposed “a skill-based fund classification” by using five spectra, each of which is shown below.
The first classification, which the authors termed “the most fundamental,” ranges from pure arbitrage — a risk-free endeavor — to risk premia.  The left endpoint is exceedingly rare in the real world (although stories about the early days of ill-fated Alameda Research talk about the presence of pure arbitrages in cryptocurrencies).

Risk premia have been the focus of attention in public markets during the last few decades, as factors have been identified and then institutionalized into strategies and products.  They stem from “persistent heterogeneities among market participants (heterogeneity in utility functions including investment horizons, presence of different types of costs and investment constraints).”  And behavioral patterns.

But, as the authors noted, the more one moves to the right along the above dimension, toward risk premia, “the more uncertainty is associated with performance.”  The spotty record of late of some widely-used factors is causing investors to question whether they should cling to previous expectations or abandon them.  Since the paper is concerned with assessing skill, the authors emphasize that “identifying the most essential systematic risk factors and correctly estimating their current risk premia represents a special skill,” beyond the simple use of factors in a set fashion.  (An additional aspect of implementation regarding risk premia comes from the use of a single factor versus multiple ones.)

In an early draft of the paper, this scale went from “quantitative to qualitative,” but was changed to the above.  Is the information that is essential to the process available in a formalized way (think prices, reported earnings, macro data, etc.) or is it part of the continuous information flow, including news events, that feeds the market?

Crucially, gathering and processing these two types of information requires essentially different skills.  Formalized information is relatively cheap to access and interpret.  However, exactly because of this reason the universe of market participants utilizing it is extremely competitive.  On the other hand, non-formalized data is hard to comprehend and apply and if implemented on a large scale, it requires extensive text mining and processing skills.

In the years since the paper was published, the explosion of activity in the areas of natural language processing and artificial intelligence have changed “non-formalized” activities in a major way.

Intertwined with the formalization question is whether information is publicly available or not.

Public information, in our terms, is information which is acquired relatively cheaply and often comes down to data vendor subscription fees.  In contrast, obtaining private information, i.e., information not readily available through public information channels, is often associated with significant ongoing expenses, be they explicit or implicit.

On the left side of the range, “private information gathering is an expensive and often technologically advanced process.”  While not mentioned, it can also be an expensive process not because of the implementation of technology but by virtue of using time-consuming human investigation techniques (not including the abuse of insider information).

On the right, decisions are “based on information already disseminated in the marketplace.”  Cheap, but hard to gain an edge of any kind that way.

Another slice involves the commonly used characterizations of top-down and bottom-up approaches.

Obviously, the two types of analysis require very different types of skills.  In reality, however, the two are often combined in some proportions, so one can rarely see their pure realizations.  But still, one of them, where the firm has more expertise, would be dominant.

Finally, there is discretionary versus systematic management.  Discretionary strategies “are supposed to be far more adaptive to changing markets and are better suited to processing hardly quantifiable information,” but are “on average less transparent and replicable” — and it is “harder to rely on past performance.”  Systematic managers “are in general less adaptive and not so suitable for processing qualitative information” (although new techniques are making that less the case).

They are fundamentally different kinds of activities:

A discretionary manager’s trade is a one-time activity in buying/selling financial instruments.  A systematic manager’s “trade” is a modification to the trading algorithm.

Creating the recipe

In the middle of each scale is the word “hybrid,” which is where many investment processes are in practice, so that leads to a question of where to locate a process relative to the two endpoints.  Presumably the positioning for an asset manager is directly related to its beliefs about markets (and about the appropriate operation of an organization in order to profit from them), the resources available to it, and the skills it can bring to bear.  If those are out of sync, the firm is unlikely to be a success.

There are layers and nuance within each of the dimensions, making a discussion about the choices potentially revealing.  Where is the firm positioned and why?  One exercise is to have members of the investment staff independently mark where they think a) the firm is positioned on each scale and b) where their competitive advantages exist within and across the five categories.  The disparities in responses may indicate different definitions or expectations (illuminating on their own), as well as more serious lapses in strategic construction.

Then consider the element of change.  Not just how the firm has migrated over time, but how it should do so in the future to adapt to long-term shifts in the market environment.

Reverse engineering the recipe

For allocators of capital, the five classifications can serve as a mapping device and a set of ideas to dig into a firm’s strategy and process in a somewhat different way, offering uncommon questions that can lead to discovery (the ultimate goal).

Recently, the New York Times published an article by Rob Copeland from his book about Bridgewater Associates.  If his portrayal is accurate, then the narrative explanation from Bridgewater about where it is positioned on these attributes is quite unlike what is actually the case.  Searching for markers of positioning and skill falls short of the sort of process attribution that you would want to have, but is a helpful step in seeing the real picture.

Just as this framework can provide insights for asset managers that want to improve, it serves as another lens through which they can be evaluated by those who vet them.

 

Given significant changes in the investment environment, now is the time to focus on innovation in your organization.  Some ideas.