| March 16, 2016
We find that earnings sentiment gives investors an edge over consensus.
In our latest research, we find that our
new earnings sentiment indicator
derived from real-time news and social media significantly outperforms consensus estimates, both in terms of Information Ratios and annualized returns.
In this study, we consider whether
sentiment data about earnings announcements
can offer investors an edge when compared to traditional consensus. Specifically, we develop an indicator that extends beyond the “hard numbers” and further considers how a company’s earnings data is portrayed in the news from a sentiment perspective.
We find that
focusing on earnings sentiment
typically yields higher Information Ratios and twice the return when compared to a strategy based on
traditional earnings consensus
. Our strategies are backtested over an 8-year period across regions and company size.
Tracking company earnings is a key element of any fundamental trading strategy. Often, a shortterm trading opportunity arises following the realization of a company either missing or beating market expectations. Gaining an edge on such information, however, is very difficult since there are literally millions of eyes on Wall Street looking at the same information.
In this study, we consider whether sentiment data on earnings related news and social media can offer investors a competitive edge. Specifically, we create an earnings sentiment indicator that we deploy as part of a unique and non-traditional thematic alpha stream.
In order to create our sentiment measure, we leverage RavenPack’s Event Sentiment Score (ESS) on news that relates to either earnings, revenues or dividends, i.e. spanning anything from changes to analyst’s estimates or company guidance, to reported figures being better or worse than expected. This allows us to move beyond the “hard numbers” to consider how earnings information is portrayed in the news from a sentiment perspective and how to weigh potentially conflicting information, i.e. positive historical results vs. negative future guidance.
Our recent research focus  has been on creating a so-called super-alpha strategy built on top of a set of thematic alpha streams . This paper considers one such alpha stream. Combining multiple alphas has proven to be an interesting angle on portfolio construction and diversification, and has been widely covered in recent literature [3, 4, 5]. In contrast to previous research, where we simply considered the average sentiment by event group, in this study we seek to exploit the more detailed information available at the category level. More specifically, we increase the sophistication of our indicator construction process by tracking the performance of any single category, and filter out the non-performing ones. Even though, in this paper, we focus purely on earnings-related news, the methodology is also applicable to other themes (e.g. Products & Services, Analysts Recommendations, Insider Transactions, etc.).
The layout of our paper is as follows: Section 2 explains our indicator construction methodology including a description of the data source that we are using. In Section 3, we analyze the performance behavior across various parameter settings in order to select the historical window size in our model. In Section 4 we show the performance we get with the constructed indicator as a signal and benchmark it against a similar strategy built on top of a standard earnings database. Section 5 summarizes our results.
Please use your business email. If you don't have one, please email us at email@example.com.
We will process your personal data with the purpose of managing your personal account on
RavenPack and offering our services. You can exercise your rights of access, rectification,
erasure, restriction of processing, data portability and objection by emailing us at firstname.lastname@example.org. For more information, you can
Your request has been recorded and a team member will be in touch soon.
High inflation has returned in developed markets after decades of lying low. In our latest paper, we show how to build an inflation-based asset allocation strategy using sentiment data and we illustrate that sentiment-based strategies outperform models that depend merely on past observed inflation values.
This year's RavenPack Research Symposium brought two intense days of knowledge sharing in London and New York, from 25 top experts in natural language processing, quantitative investing and machine learning. Together, we explored how firms can leverage new language models to generate alpha, better manage risk and respond to calls for more sustainable investment practices.
Human capital is at the heart of value creation. Our latest research demonstrates how unprecedented workforce insights, sourced from over 200 million job postings, can generate more alpha.