Earnings Intelligence: your questions answered

The RavenPack Earnings Intelligence Factors are a collection of daily scores derived from news, transcripts, insider transactions, and earnings dates, that provide condensed insights into market-moving earnings events. Backed by quantitative research, these factors are designed to facilitate efficient decision-making for investors by offering valuable information on companies' outlook during earnings season.

The RavenPack's Earnings Intelligence Factors facilitate efficient decision-making by condensing vast earnings insights into actionable signals. Research has demonstrated robust performance across regions, capitalization, and holding periods.

The evidence of effectiveness comes from back-testing results:

Strong performance across various universes and holding periods (daily to monthly).

Combining factors improves information ratios (e.g., U.S. Mid/Large-cap: 1.2, U.S. Small-cap: 2.6 over two weeks).

Even for quantitative investors who rely heavily on data-driven models, RavenPack Factors offer several distinct advantages: they reduce the need for extensive data cleaning and pre-processing, freeing up your time and resources to focus on model development, back-testing, and optimization. The additional insights offered by these factors can improve the overall accuracy and performance of your quantitative models.

RavenPack's Earnings Intelligence Factors address a major challenge for discretionary investors: efficiently incorporating and analyzing earnings data. They streamline the process by simplifying complex information into a single, actionable score. Here's how:

The factors cut through the noise by digesting vast amounts of earnings-related data, including news articles, transcripts, insider transactions, and even earnings dates. This translates into a simple signal that complements your existing fundamental analysis by offering quantified data points
You can leverage them to verify or refine your existing hypotheses. This frees up valuable time, allowing you to focus on interpretation and strategic decision-making.
The RavenPack Factors Library goes beyond pre-built options and offers a custom library. This allows you to explore vast datasets and uncover hidden trends through your own factor creation.

The back-testing covered over 15 years of data, providing a robust evaluation of the factors' performance across different market cycles. The factors were tested across various regions (U.S., Europe, Asia-Pacific) and market capitalizations (Mid/Large-cap, Small-cap) to assess their effectiveness in different market segments. The factors consistently achieved notable Information Ratios over the back-testing period (e.g., U.S. Mid/Large-cap: 1.2, U.S. Small-cap: 2.6 over two weeks) and they exhibited complementarity, meaning the combination of factors outperformed the individual performance of each factor. This diversification helps to mitigate risk and potentially improve portfolio performance.

The factors are updated daily at various snapshot times, including one hour before the market opens or closes in New York, Paris, and Tokyo. These updates occur periodically to ensure that the factors reflect the latest earnings-related information and insider activity, providing traders with up-to-date insights at common market entry points

RavenPack ensures data quality and accuracy through rigorous data collection, cleansing, and validation processes, leveraging advanced technologies and expertise in natural language processing and sentiment analysis.

Sentiment analysis on news and transcripts is performed using advanced natural language processing techniques to extract sentiment analytics, and other relevant insights from textual data.

RavenPack analyzes a comprehensive dataset of global insider transactions from over 60 sources of regulatory information, covering over 550,000 insiders across various positions in 50 countries and 50,000 companies. To identify the most relevant signals, the analysis focuses on purchases and sales of company stock, excluding option exercises and other non-trading activities. Transactions are further filtered based on:

Insider Level: this categorizes insiders by their relative importance within the company, from top executives (CEO, CFO, etc.) to lower-level executives.
Transaction Significance: this captures regular buying and selling activity, excluding non-intentional or mechanical transactions such as stock awards, option exercises, tax-related transactions, remuneration, and share plan purchases.

The RavenPack Factors library goes beyond pre-built options with a custom library, which lets users explore vast datasets and uncover hidden trends through their own factor creation.