Alternative Data and the Hunt for Alpha: a Rich Man’s Game

Maria Adamjee is CEO of  Megalodon Capital   

While The terms “Alternative Data” and “Alpha”—the industry equivalent of “dude” or bro”—are ubiquitously  peppering Financial Markets discussions as fund managers seek differentiators in a highly competitive hedge fund market. Rarely does one encounter investment or trading content that can resist flaunting the term—but in practice, leveraging Alternative Data, especially to achieve alpha, is much more difficult than one might imagine, necessitating in the end, a rethink in regulatory policy.

To understand why, it’s important to understand precisely what everyone is (supposedly) talking about. “Alternative data” refers to any data set that is considered outside of standard market tick data. Although the term was typically used to denote ESG (environmental, social and governance) content, it has evolved to include any source of information that can contribute to idea generation. ‘Alpha’ in trading is the access of returns on an investment measured against the movement of the market as a whole. Hedge Funds also use the word “Alphas” in reference to quantitative models which are built to find trading patterns in new data sources. 

Because fund managers suffer from “Fear Of Missing Out” when it comes to Alternative Data, it is constantly at the forefront of their minds. The adopters of “Alt Data” today are a small group of about 20 firms, mostly in the equity stat arb space and mostly in the US, that have the capabilities and resources to evaluate a wide array of raw alterative data sources. That number is gradually growing, as investment firms are increasingly seeking out sources of Alternative Data to deliver Alpha. Still, due to the associated costs and practical challenges of data sourcing, this option is not feasible for all funds.

Rhetoric like “generating alpha from alternative data” gives the false impression that alpha in the markets is infinite and ubiquitous. In practical application, this is far from the case. The search for material returns requires resources, capital and time and is a multi-step process. 

Step 1: Data Procurement

Fund managers hire teams of specialized talent (“Data Hunters”) who, as their name suggests, seek and identify unique sources of data. Many of these data sources are not called “Alternative Data providers”; instead, they are often companies that are collecting data for their own purposes. The hunters will approach firms to understand how the data can be useful for their Fund’s investment strategy and how it can be classified. 

Once they determine that the data is pertinent, the hunters compose short theses to explain how the data can be tied back to specific equity names and why it would be useful for the fund manager. They also look at how far back the data can go so they can back-test against a variety of market conditions, and determine whether the data can easily integrate into the next phase: Validation  

Step 2: Data Validation

Data Scientists scour the content and determine if any patterns in the data can be correlated to either sector level or single name predication. This assessment involves back-testing data against live market data to understand the frequency and strength of the alpha signals.

Once the tests certify that the data has value, it is presented to portfolio and investment managers along with back-test results and the investment thesis before the final phase:  Integration.

Step 3: Data Integration

Once the Alpha is determined, fund managers need to integrate this data into their existing models. Portfolio Managers create mathematical functions that are coded into trading algorithms that make automated decisions to buy or sell a particular stock or sector based on the data’s suggestions. 

__________

The handful of sophisticated funds that currently have procedures in place for data sourcing have another blind spot amongst all consumers of data: price transparency.

The new SEC administration, under which Brett Redfearn is the director of trading and markets, has taken up the mantle of challenging the exchanges’ market data fees, questioning whether the fees are aligned with the exchanges’ mission of charging fair, reasonable and non-discriminatory rates. However, unlike ticker data, there is no such evangelist in the Alternative Data space. Alternative Data providers today set prices based on the fund managers’ Assets under Management (AUM). If they expect that a manager will make a 10 to 20 times multiple return on a data investment, they will assess how much AUM a manager will deploy to their Alt Data and then provide a price.

This model favors firms that have the most AUM and are therefore best equipped the deploy diverse Alternative Data Strategies.  

The Alt Data space isn’t fair or transparent. There are few rules for governance and almost no support for mid-sized to small funds that need to infuse data into their strategies to stay relevant. For professional traders, finding new methods of improving active return on investment (Alpha) is more critical than ever. While Alt data has the potential for big returns, it is currently reserved for the few.

The easy obtainability of public tick data and technology, together with more intense competition, makes cost reduction and price transparency in the Alt Data market necessary. Firms with a capital edge are sourcing all the unique data and squeezing relevant Alpha out of it, diluting it so drastically that by the time it reaches the next management tier, the data is almost completely devoid of value. This data arms race is reminiscent of the “need for speed” in the early 2000’s when HFT firms with enough capital to get close proximity to the exchange and route orders quickly over custom infrastructure had a Lance Armstrong-like advantage – until regulators stepped in to level the playing field. 

A handful of data-brokers and consulting firms have intervened to bridge the gap by scouting and sourcing interesting data sets on behalf of investors; valuating the data for breadth, length of history, economic intuition, survivorship bias, completeness, and accuracy; and building models by applying the same sort of rigor a top quant firm would apply to the analysis. 

While buy-side firms would like to adopt alternative data as part of their processes, relatively few have made more than nominal progress.

About The Contributor: admin

Get involved!

Get Connected!

Come and join our community. Expand your network and get to know new people!

Comments

No comments yet