Understanding News Flow - Country Distribution and Seasonality

To benefit most from the information edge offered by RavenPack, it's necessary to understand the cross-sectional and time series patterns of the RavenPack News Analytics database.

Currently, RavenPack offers two products in the market place including the Dow Jones and Wired Editions. This study focuses on the news flow distribution of each product as well as on the combined set.

News Flow Differs Significantly Across Countries

• The US accounts for 42% of total news flow within the Dow Jones Edition.
• The Top 10 countries ex. US together account for about 37% of total news flow.

News Flow is Characterized by Strong Seasonal Patterns

• Considering the business week, equity news tends to peak on Thursdays, and having the least news on Mondays and Fridays.
There is hardly any equity news on weekends.
• In the US, most news flow occurs in October and the least in December.

De-trended News Flow Can Help Improve Correlations

• Removing seasonal effects reduce the impact of cyclical components on aggregated sentiment, and improves on the ability to capture structural changes.
• De-trending increases the contemporaneous correlation between the S&P 500 and our aggregated sentiment measure from 20% to 38%.

Country Distribution of Total News Flow

RavenPack has partnered with Dow Jones, a global leader in financial news, to develop the RavenPack News Analytics - Dow Jones Edition. The news sources include Dow Jones Newswires, regional editions of the Wall Street Journal, and Barron's. For back-testing purposes, news analytics data is available since January 2000.

RavenPack has also partnered with other leading news organizations including the Associated Press, PRNewswire, Wall Street Horizon, MidnightTrader News, the Canadian Press, among many other publishers to develop RavenPack News Analytics – Wired Edition. For the Wired Edition, news analytics data is available since January 2007.

In this report, we take a closer look at cross country news flow distributions as well as the main seasonality patterns of both of these datasets. In particular, we provide a comprehensive overview of the RavenPack News Analytics database in terms of country distribution of: (1) total news flow, (2) significantly relevant news flow (using Relevance Score≥90), (3) highly relevant news flow (Relevance Score=100), and (4) the most novel and highly relevant new flow (Relevance Score=100 and Event Novelty Score=100).

Relevance is a score between 0-100 that indicates how strongly the company relate to the underlying news story, with higher values indicating greater relevance. See Appendix A for more details on the Relevance Score. The Event Novelty Score is a score between 0 and 100 that represents how "new" or novel a news story is within a 24-hour time window across all news stories. The first story reporting a categorized event about one or more companies is considered to be the most novel and receives a score of 100. See Appendix A for more details on the Event Novelty Score (ENS).

The result for the Top 10 countries in terms of total news count in the Dow Jones Edition is presented in Fig 1 and Fig 2 (please refer to Appendix B for the results on all countries). The Dow Jones Edition has news coverage for 93 countries and includes on average about 16,500 news records per day. Generally, news flow tends to concentrate on a few major economies. Fig 1 shows the news flow distribution in the Top 10 countries . These countries are, in the order of rank, the US, UK, Japan, Canada, France, Germany, Australia, Switzerland, India, and Hong Kong. Together they account for 78.9% of total news flow. As expected, US news flow dominates other countries with 42.1% of news records with a daily average of almost 7,000 records. The Top 10 ex. US together account for about 37% of total news flows. Countries that are considered actively covered...