News Sentiment Everywhere - Trading Global Equities

RavenPack | July 09, 2018

This white paper demonstrates how news sentiment can be a valuable input to alpha generation.

News sentiment is everywhere and in order to maintain an edge in the marketplace, asset managers are to a larger extent turning to unstructured content for alpha creation, using NLP and text analysis techniques.

In addition, more and more managers are expanding their mandate, trading global portfolios, to ensure more scalable strategies. This white paper demonstrates how news sentiment can be a valuable input to such process.

We are using country-level long/short equity strategies, focusing on large- and mid-caps.

News Sentiment Everywhere: High-Level Findings

  • Returns are positive in 41 out of 49 countries, with Information Ratios (IRs) greater than 1.0 in one out of three countries tested.
  • When combining our strategy into regional portfolios, the added diversification helps to achieve Information Ratios of 3.0 or higher in 3 out of 5 regions, including North America, Europe, and Asia Pacific. News Sentiment
  • We ensure maximum diversification by combining our strategies into a global portfolio, with Information Ratios as high as 4.8 for 1 day holding period, and 2.6 for a 1-week horizon.

News Sentiment

Strategies are based on the recently introduced Sum Excess Sentiment Indicator (SESI) that adjusts for daily news sentiment bias . It can be calculated using RavenPack’s Self Service Data Platform.

Commentaries on the News Sentiment Everywhere White Paper

by Peter Hafez, Chief Data Scientist, RavenPack

With the emergence of alternative data, the alpha-landscape is in the process of getting seriously disrupted. The winners will be the firms that understand how to embrace and take advantage of all of the new opportunities arising from data abundance.

Alphas have become much faster than before, increasing the pressure on asset managers to find new ways of scaling their strategies to maintain their performance on the same level of AUM. While new data sources are helping to alleviate this issue, investors have, to a larger extent, been expanding their mandates from trading locally to trading globally.

Unfortunately, applying alternative data on a global scale can be challenging because it is not easily available, and often lacks coverage. Furthermore, many datasets do not come with the depth of archive required to perform a proper backtest, or they may lack point-in-time sensitive scoring or tagging. Within the alternative data landscape, news analytics is a real positive outlier.

To showcase how news analytics can be used as a source for creating globally scalable and diversified alphas, in our latest research we created a simple sentiment signal, applying it on a global scale. Specially, we considered RavenPack’s Event Sentiment Score (ESS) to create a Sum Excess Sentiment Indicator (SESI), which was used as an input for market-neutral strategies at country-level. Finally, these were combined into regional and global portfolios to achieve further diversification benefits.

News Sentiment Everywhere Video

Watch the News Sentiment Everywhere presentation highlights from our recent Big Data and Machine Learning event in London here .


News Sentiment

Introduction

In recent years, continuous advancements and the adoption of Information Technology (IT) have resulted in ever faster exchanges of information across the globe and a significant increase in the production of digital content. All of this has helped fuel the Big Data and Artificial Intelligence (AI) revolution that is affecting almost all aspects of modern society.

The impact on the financial industry has already been significant. The influx of new data sources, i.e. alternative data, has provided investors with a considerably larger array of investment opportunities, and has made it substantially easier and faster to operate on a global scale. Accordingly, alternative data is quickly becoming the new frontier for alpha generation.

As a consequence, the alpha-landscape is in the process of getting seriously disrupted. Using alternative data, the number of potential alpha streams has increased significantly. However, each alpha has become much faster than before, increasing the pressure on asset managers to find new ways of scaling their strategies to maintain their performance on the same level of AUM. While new data sources are helping to alleviate this issue, investors have, to a larger extent, been expanding their mandates from trading locally to trading globally.

Unfortunately, applying alternative data on a global scale can be challenging because often it does not exist, is not easily available, or simply lacks coverage. Furthermore, many of these datasets do not come with the depth of archive required to perform a proper backtest, or they may lack point-in-time sensitive scoring or tagging. Within the alternative data landscape, news analytics is a real outlier. Available products have greatly matured over the years, making them easier to consume, and the nature of the underlying sources provides unique insight into global information flows.

Specifically, RavenPack Analytics (RPA) provides an extensive history of news analytics data going back to 2000 with a global coverage of more than 45,000 companies across 143 countries, characteristics making it a natural candidate to cater scalable and diversified alphas at a global scale.

As part of this study, we showcase the value of news sentiment as a source of alpha on a global scale. To investigate the responsiveness to news across different markets, we create single stock strategies across 48 countries and analyze their performance. Since the sparsity of the sentiment signal in some markets can be detrimental to performance, we extend the analysis by combining country portfolios into regional and global portfolios to ensure maximum diversification and risk-adjusted performance.

This research is related to well established academic studies that have tried to analyze how traditional factors and quant strategies perform in different countries, across regions and globally: Fama and French [1] show that a value premium is present in all regions and is generally more marked for smaller stocks, while strong momentum premium is found in all regions but Japan. Similarly Asness, Moskowitz and Pedersen [2] find that Value and Momentum factors work everywhere, across different regions1 and asset classes, with a strong correlation structure among strategies built on the two factors. Even more relevant to our research, other studies have shown that news analytics can improve traditional strategies in specific countries or regions.

Among others, Burke Lau [3] shows that after conditioning on companies with low media attention, it is possible to observe a momentum premium also in Japan. This can be justified by the fact that investors tend to overreact to stocks receiving higher media attention, which translates into a mean reversion in the following month for such companies. More recently, Ada Lau et al. [4] show that in Japan “no news is good news for reversal players”, as they find that stocks with no news coverage exhibit a strong short-term (1-week) reversal, probably justified by the fact that investors might overreact to “non-news-based information or spurious price movements”. Additionally, using News Analytics has proven to be useful in reducing divergence risk in traditional pairs trading strategies in U.S. and Europe, by excluding pairs that exhibit price divergence supported by news as shown in Qu et al. [5].

In this study, we adopt a more organic approach by focusing solely on sentiment strategies — based on the recently introduced Sum Excess Sentiment Indicator (SESI) — and by analyzing their performance on a country level, as well as regional and global scales. The paper is organized as follows: Section 2 describes both the pricing and the sentiment data used in our research. Section 3 follows with a description of the methodology implemented to build country level, regional, and global portfolios based on sentiment. In Section 4, we present results for the different strategies both in terms of excess and factor-neutral returns. Section 5 presents our conclusions.



By providing your personal information and submitting your details, you acknowledge that you have read, understood, and agreed to our Privacy Statement and you accept our Terms and Conditions. We will handle your personal information in compliance with our Privacy Statement. You can exercise your rights of access, rectification, erasure, restriction of processing, data portability, and objection by emailing us at privacy@ravenpack.com in accordance with the GDPRs. You also are agreeing to receive occasional updates and communications from RavenPack about resources, events, products, or services that may be of interest to you.

Data Insights

Read More