Outline for an Alternative Data Hedge Fund Strategy

DataScrum
DataScrum
Jun 18 · 4 min read

Introduction

Imagine you wanted to go ahead and create your own hedge fund strategy based on alternative data today. Let’s give it a name for the sae of this example and call it ‘AltCap’ . You would then have to put together a pitch to LPs (limited partners) to fund such an enterprise. What would the outline of this pitch look like?

Below are key building blocks and a road map how to put together such a fund:

Elevator Pitch

We propose a trading strategy that relies on using alternative data, such as web traffic of ‘pure-play’ publicly listed companies, including online retailers, booking systems and publishers to beat the market consensus forecast of earnings of such companies.

Problem

The market consensus of revenue and earnings announcements of publicly listed companies today relies primarily on the following types of financial data:

Fundamental research — fundamental analysis of publically available financials and news by specialist research houses, buy-side and sell-side analysts

Earnings calls — analyst gather additional information such as sentiment from regular earnings calls to adjust up or down their consensus estimate

Expert networks — companies such as GLG tap into partially non-public information and sell it to the buy side to provide them an edge

All of these approaches have limited value as (i) the company financials often lag behind the actual performance, (ii) earnings call sentiment data have shown limited value, and (iii) the practice of using expert networks is subject to regulatory scrutiny and carries legal risks due to insider information laws.

Over time markets are bound to become more efficient as most market participants rely on the same data. As a result, market participants are increasingly relying on various categories of alternative data to assess the fair value of the publically traded assets. The major categories of such data are outlined below.

Solution

An alternative data hedge fund solution would have the following key steps, or a roadmap :

Revenue and earnings estimation using alternative proxy data — we will estimate company revenues / earnings using proxy with novel data sets, such web site visit data, assuming relatively stable bounce and conversion rates, or transaction data (debit cards) and satellite data (foot traffic) as a suitable proxy for actual financials while building a database of such data (Step 1)

Adjustment of estimates — we need to normalise and adjust our estimates. For instance, as the third-party traffic estimates are inaccurate, they will need to be revised upwards / downwards using a set of proprietary techniques (Step 2)

Model calibration and backtesting — we use past data to calibrate and backtest the model parameters and obtain a non-consensus estimate (Step 3)

Trading strategy — only given successful back tests for a sufficiently large data set we will deploy the following trading strategy (i) if the non-consensus (NC) estimate is higher than the market consensus estimate (and sufficiently far away), we take a long position, otherwise (ii) we take a short position (Step 4)

Fundamental review — the data sources we will need to go through fundamental analysis and review and continuously updated to obtain more accurate estimates (Step 5)

Market and Competition

What about competition? Surely we are not the only kids on the block attempting to deploy this strategy. It is worth nothing that while alternative data are used primarily in the US equities and global commodities markets, AltCap will try to find new markets and opportunities in:

Horizontal, geographic long tail— any country new with a public stock market listing online-only companies. For example, we can hypothesise this strategy is not yet widespread in EMEA and Asian markets

Sector long tail— in each new geographic market, we can find new non-traditional vertical sectors beyond the obvious ones such as e-commerce to identify new trading opportunities

This was easy! Please send us any feedback at team@datascrum.co.uk if you agree or disagree with our plan! We are a growing community at DataScrum of data scientists and data science enthusiasts. You can follow us on twitter or MeetUp as well. We would love to hear back from you!

Resources:

Web Traffic Prediction Accuracy — by Rand Fishkin of SEOmoz

What is Alternative Data — by DataScrum

BattleFin.com — alternative data trading competition

DataScrum Meetup — an alternative data Meetup

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade