Using AI to Uncover Signals in Corporate Event Data

  • By Neuravest
  • July, 23

Wall Street Horizon’s forward-looking corporate event data feeds provide financial services professionals with clear and reliable views of potentially market-moving developments. In response to analysts, portfolio managers, traders and chief investment officers becoming more quant-savvy, Wall Street Horizon leveraged the Neuravest Data Analytics Services (DAS) platform to validate the predictive nature of its corporate events data and enable fund managers to apply actionable signals derived from this data to their investment strategies via model portfolios and smart data feeds.

Read along as Christine Short, Vice President of Research at Wall Street Horizon, and Eric Davidson, Executive Vice President, Strategic Partnerships at Neuravest, describe the relationship between the two firms and how the right approach can result in powerful validation for complex datasets.

Q: Give us a brief overview of Wall Street Horizon. What is the company’s background and what does it do?

Christine Short: Wall Street Horizon was founded in 2003 by Barry Star. The company launched in response to a fundamental problem: the highly manual and repetitive process of compiling corporate event data, which is critical for options traders and market makers. For the past 17 years, Wall Street Horizon has been exclusively focused on forward-looking corporate event data. We are the gold standard for accurately tracking, curating and presenting event data to institutional clients worldwide. Our ultimate goal is to ensure that our clients don’t miss out on events that will cause volatility in their portfolios. We are based in Woburn, Massachusetts and today have approximately 50 employees around the globe.

Q: What challenges do you experience when explaining the value of your data to fund managers? How do you overcome them?

CS: The most common response from skeptical funds is that they already have a data terminal or that their current solution is “good enough.” This gives us an opportunity to highlight that our accuracy and reliability are superior to any of the big box platforms. If these funds want to outperform their peers, they need to go beyond the terminal and take steps to ensure their corporate events intelligence is verified and up to date. When funds make investment errors due to bad dates, they often are not aware something better is available.

Eric Davidson: Investment managers are always looking for ways to outperform. Many traditional data sources have become commoditized and no longer provide an advantage – which means PMs are looking for analytical and informational advantages in ways that they may not have in the past. Wall Street Horizon provides a unique view into corporate event data that isn’t readily apparent to many investment managers, and communicating the predictive power of this data is essential to their success as a firm.

Q: Can you talk a bit more about your datasets? How do you get them? For which ones do you leverage the Neuravest DAS?

CS: We currently cover more than 40 event types – things like earnings announcements, shareholder meetings, corporate actions, dividends, conferences – across approximately 9,000 publicly traded companies worldwide. Our data is primary sourced, meaning we get it directly from these companies via phone and email communication, company-issued press releases, shareholder calls, corporate filings, company web sites, company videos, company RSS feeds and more.

We use the Neuravest DAS for our DateBreaks dataset, which captures earnings date revisions. This data is inspired by what we call corporate body language, which consists of non-verbal corporate communications that can act as a tell for a company and its financial health. Recent academic research has shown that tracking changes to earnings announcement dates can help investors generate additional alpha or mitigate risk in their portfolios. Neuravest provides important third-party validation for this concept by using our data to develop predictive investment signals and model portfolios that can be brought to market quickly.

ED: There is an ocean of content and information sources out there – and thus the question becomes, what’s worth pursuing as a portfolio manager or trader? Wall Street Horizon tapped Neuravest to showcase to PMs and traders that their dataset is robust, high-quality and predictive. The Neuravest value-add is to help Wall Street Horizon turn this data into a competitive advantage for fund managers by proving it empirically. In addition, data providers are eager to distribute their insights to the larger market of investment managers – not just the large quantitative hedge funds. The Neuravest platform unlocks that distribution channel for data providers, growing their addressable market by generating model portfolios and smart feeds that are easily licensed or ingested by investment managers – without the need for quantitative staff.

Q: What goes into this validation process?

ED: As Christine mentioned, our work has been on the DateBreaks dataset – specifically, to illustrate that when a company confirms an earnings date, and then subsequently changes it to a later date, it is likely that the company is preparing to share bad news. Conversely, when a company accelerates an earnings date release, the company is likely to share good news. This is an intuitive dynamic and it’s no surprise the signal is meaningful – companies are excited to announce the positive and delay the negative.

While the Wall Street Horizon signal by itself is meaningful and predictive, we leverage machine learning to enhance the signal and incorporate additional technical and fundamental factors. The result is a dynamically adjusting model portfolio with Wall Street Horizon signal drivers at its core. Ultimately, the Wall Street Horizon signal helps drive a portfolio that outperforms a given benchmark. Neuravest validated this in-sample, out-of-sample and now for more than a year in forward-testing.

Q: Tell us more about these portfolios. How does Neuravest build them using identified signals?

ED: The first step is to make data “AI ready.” In Wall Street Horizon’s case, they have model-ready data available at the outset – ready to be analyzed with algorithms and by quantitative researchers. Neuravest then validates the data – with both descriptive studies to understand the shape of the data and predictive analytics – to find signals in the data. We follow that with feature identification and extraction, model training and in-sample and out-of-sample validation. There are quite a few steps in this methodical and rigorous process, whereby our researchers work in conjunction with our infrastructure and solutions stack to construct portfolios worth paying for. The final step is launching the portfolios into production – and end-to-end solution.

An oil refinery can serve as a useful analogy for what we do, as it really is akin to a manufacturing process. Data is the oil and machine learning is the combustion engine. We remain transparent with the data supplier through every step of the process, including regular check-ins and deliverable updates as well as knowledge transfer – particularly early in the process – as we very much like to incorporate the domain expertise of the data partner.

Q: What does this mean for fund managers? How has this impacted Wall Street Horizon’s business?

ED: A large portion of investment managers don’t have quantitative and data science talent on staff, yet they need the insights that come from new data sources and non-traditional ways of analyzing information to stay competitive. Neuravest addresses this pain point in a range of ways, including validation through both historical and forward testing. We also provide an environment for portfolio managers and traders to “try before they deploy” by giving them access to live model portfolios and smart feeds. With this is transaction-level data – and what we call “last mile implementation” – investment managers can seamlessly incorporate these alpha sources into their research and portfolio construction environments.

CS: Several clients and prospects have approached us looking for signals, as we don’t provide them, and we have been able to point to Neuravest’s work as proof of the predictive power of DateBreaks. While we have full confidence in our data and believe its utility is self-evident, it is great to be able to point to concrete, third-party validation when funds ask us to make our case.

ED: In addition, Neuravest currently has a number of buy-side customers that have interest in utilizing the Wall Street Horizon dataset, and we expect to convert them into customers for both Wall Street Horizon and Neuravest soon.

Q: Tell us more about the relationship between the two firms. How did it come about and what is the ongoing relationship like?

ED: We’ve been working with Wall Street Horizon for several years, ever since we approached them with the idea of providing quantitative services and tools to the next tier of practitioners, meaning those just getting started with systematic approaches.

CS: We have always had a strong presence in the quantitative space, but our relationship with Neuravest has been marked by smooth processes and effective communication. We sometimes correspond on small questions or strategic brainstorming, but for the most part, we simply hand over the data and allow them to go to work.

Q: What’s next for Wall Street Horizon? What future initiatives do you have planned, and how will Neuravest be involved?

CS: We’ve confidentially shared our product roadmap with Neuravest and look forward to continuing to work with them. As new market data providers continue to enter the space and more traditional investors seek additional ways to gain an edge through quantitative processes, we believe there will be a greater need for the kind of third-party validation Neuravest provides and encourage market participants to watch this space.

ED: For those who are interested, we provide a real-time feed of the Wall Street Horizon portfolio on the Neuravest website. We are in production with the Wall Street Horizon model portfolios now, and that means a tighter partnership and collaboration with their team. When you think about business strategy in general, all roads lead to people and, in particular, talented people, which Wall Street Horizon has in spades. Neuravest is very excited to work with their team as we build out this distribution channel. More broadly speaking, we stand ready to help data providers stand up new offerings and provide the necessary validation, no matter what form the data takes.




Have a media inquiry or a topic you’d like to contribute to our blog?