Alexandria’s research and development is grounded in the kind of interdisciplinary study that makes possible breakthroughs across specialties.
Risk Systems That Read
Summary: With Risk Systems That Read, Northfield measures how the present is different from the past and, therefore, how the near future is also likely to be different from the past — something missing from nearly all financial models.
Empirical Research Partners: Robot Newsreaders
Summary: Computing net sentiment for economic topics helps investors see what is directly driving episodes of panic or euphoria, allowing them to position their portfolios over 12 month horizons.
The Case for SenTMap
Summary: SenTMap is for investment professionals who are overwhelmed with financial news, fear they may not be taking full advantage of it, and don’t have a quick and efficient research environment for exploring its effects on their holdings.
Heart of the Matter
ACTA’s Contextual-Intelligence Engine
Summary: The ACTA engine is a set of proprietary algorithms that assess the sentiment of unstructured information – such as high-value financial news feeds. The output, scored for negative or positive sentiment, is then delivered as real-time signals or a set of indices to clients, who use such contextual-intelligence to predict the outcome of their actions and to assess the risks involved.
Maximizing Returns with Alexandria Event Tags
Summary: Alexandria\’s ACTA engine identifies 23 events within news. Identifying these events improves the return profiles for news based strategies for one month and three month intervals. Factors were weighted based on IC\’s as derived from back-tests of the individual news events.
Sentiment v. Momentum
A Comparative Factor Analytsis
Summary: Net Sentiment outperforms 9M Momentum over the research period when sentiment derived by Alexandria from Dow Jones news articles are aggregated to create net sentiment metrics for a universe of stocks.
Sentiment Applied to Multi-Factor Models
Summary: Monthly sentiment may be included in the portfolio construction process to increase returns:
- As a stand-alone factor, monthly sentiment [1M_Sent] may be used to select securities that outperform the universe and equal-weighted benchmark.
- When added to multi-factor models, sentiment may be used as a qualitative screen to delay purchases with negative sentiment and sales with positive sentiment, improving portfolio returns and reducing risk.
- Portfolio returns improve when sentiment is added as a factor within a multi-factor model.
- Incorporating sentiment as a factor and screen produces portfolios with the most favorable results among the applications tested.
Sentiment Assessment for Trading Applications
Summary: Identify trends between sentiment and price by importing Alexandria’s sentiment scores into trading applications. Convert Alexandria’s sentiment scores into effective indicators – including a Relative Sentiment Index (as below), Moving Averages, MACD and more.
Sentiment for Risk Management
Summary: News based sentiment can be used to delay the timing of buy and sell decisions by active investment managers. By creating daily sentiment metrics derived from news reported in the Dow Jones Global Newswire, managers have the ability to monitor positions sent to trading desks for execution. Using this information, managers can delay the timing of buy decisions for companies receiving negative news today until negative sentiment diminishes overall, and vice versus for companies with positive news.