impact on their position or portfolio. Also, news events on earnings, product recalls,
layoffs, stock or credit ratings, and many other categories can be precursors to changes
in volatility of securities. Structured news or ‘‘news analytics’’ enable traders to get a
head start by acting in advance of these changes.
The underpinnings of news analytics
Being able to express news stories as numbers permits the manipulation of everyday
information in a mathematical and statistical way that allows computers not only to
make decisions once made only by humans, but to do so both faster and more efficiently.
Since market participants are always looking for an edge, the speed of computer
connections and the delivery of news analytics, measured in milliseconds, have become
an essential part of investment decision making.
News analytics are used in financial modeling, particularly in quantitative and
algorithmic trading. They are usually derived through automated text analysis and
applied to digital texts using techniques from natural language processing and machine
learning such as latent semantic analysis, support vector machines, and Bayesian
sorting, among other techniques.
News data are delivered in a variety of formats, often as machine-readable XML
documents or delimited files. Analytic data include numerical values, tags, and other
properties that represent certain aspects of the underlying news stories. For backtesting
purposes, having historical news analytics data is key. Usually, historical data are
delivered via flat files, while live data for production purposes are processed and
delivered in milliseconds through direct data feeds or APIs.
News analytics work well with most investment strategies and can even help to
improve investment and trading performance. To many financial participants, news
analytics are considered a relevant, novel, and even critical input to their decision-
making process. From a risk management perspective, firms can look for measures
to start trading, halt trading, widen spreads, or hedge with other instruments based
on how they anticipate the market will react to news. Generally, the value add of
automated news analysis has caught the interest of professionals and researchers,
who have shown promising results focusing on predicting stock price direction,
volatility, and trading volume.
Quantcentration and news
Trading model inputs are traditionally derived from company fundamentals and market
data. Traditional quantitative factors are crowded and performance has been degraded
due to ‘‘quantcentration’’—whereby most firms use the same type of data in their
models. This is one of the reasons investment models have performed poorly in recent
years.
Independent of traditional factors, news analytics are a unique source of explanatory
and predictive input. They include structured information and signals that create new
trading opportunities on both scheduled and unscheduled news events. Today, these
data are used to power a number of applications ranging from high-frequency trading
systems requiring low-latency inputs, to risk and asset management applications
requiring factors whose time resolution may be daily, weekly, and monthly.
312 Industry insights, technology, products, and service providers